Unsteady MHD transport of a tangent-hyperbolic ternary hybrid nanofluid over a permeable stretching surface with variable thermal conductivity and nonlinear radiation Authors: Abanti Datta, Amitava Dutta, Arka Banerjee, Sayantan Mukherjee Published: 2026 ABSTRACT AbstractThis work is a numerical exploration of the unsteady MHD transport of a ternary hybrid nanofluid of tangent-hyperbolic flow over a permeable stretching surface. The working fluid consists of Cu, and nanoparticles dispersed in ethylene glycol. The model includes nonlinear radiation, variable thermal conductivity, Joule heating, and Darcy-Forchheimer porous effects. The governing equations of partial differential equations are reduced to similarity equations of ordinary differential equations and solved by MATLAB BVP5C. Simulations show that an increase in the Weissenberg number and Forchheimer parameter decreases the velocity field because of the increase in elastic and inertial resistance and the energy dissipation elevates the fluid temperature. The variable thermal conductivity will greatly enhance thermal gradients near the wall and increase the performance of heat transfer. Moreover, increased concentrations of nanoparticles enhance skin-friction coefficients, Nusselt and Sherwood numbers. The ternary hybrid nanofluid is better than mono and binary systems due to synergistic interactions of the nanoparticles that promotes thermal conductivity and mass diffusion. The findings indicate that the Nusselt number decreases up to 92.14% with increasing magnetic parameter and decreases to about 95.93% with increasing thermal radiation, which shows a high level of suppression of heat transfer. Conversely, the stretching parameter augments the skin friction and mass transfer by approximately 42.48 and 11.47% respectively, whereas thermophysical parameters augment the Nusselt number by up to 286.9%. Comprehensively, the numerical analysis shows, quantitatively, the competing interplay of electromagnetic, thermal and diffusive processes in dictating transport behaviour in ternary hybrid nanofluid flow. These results demonstrate the possibilities of tangent-hyperbolic ternary hybrid nanofluids in the advanced cooling and thermal management processes with intricate rheological and electromagnetic interactions. The soundness and correctness of the numerical framework is validated against existing literature.Keywords: Hybrid nanofluid, tangent-hyperbolic model, Unsteady MHD, Nonlinear radiation, Darcy-Forchheimer effects Introduction Fluids with non-Newtonian rheology sought attention of the researchers due to its increasing demand in diagnostics, drug delivery, polymer processing, and advanced chemical systems [1–3]. While the Newtonian fluids are known for its invariable viscosity over the ranges of shear rate, the viscosity of the non-Newtonian fluids varies with the same. While most of the studies on non-Newtonian rheology are carried out on pseudoplastic, dilatant, Carreau, yield pseudoplastic and viscoelastic models, the tangent hyperbolic fluid (THF) model is underexplored so far. The THF model uses a hyperbolic tangent function to describe a non-linear stress–strain relation capturing both the viscous and elastic effects under varying shear conditions. This feature makes it suitable for modelling polymer and biological flows [4–6]. Nanofluids are typically suspensions of metallic, oxide-based, or carbon-based nanoparticles (1–100 nm) in a base fluid like water, oil, ethylene glycol etc. The nanoparticles are added to improve the heat transfer performance of the base fluid. Mono-nanofluids are base fluids containing a single kind of nano particles, whereas in hybrid nanofluids more than one kind of nanofluids are dispersed in a base fluid. The term ternary nanofluid indicates that the types of nanoparticles are three. The ternary nanofluids are known to exhibit improved thermal conductivity, better stability, and greater heat transfer efficiency compared to mono and binary nanofluids. Their wide used in electronics cooling, heat exchangers, renewable energy systems, and biomedical device has grown interest of the researchers in last two decades mostly. Entropy generation, particle shape effect and advanced heat transport model studies on nanofluids are carried out by several researchers. Interested authors can be referred to Boujelbene et al. [7], Alqawasmi et al. [8], and Jat et al. [9]. The study of Mahboobtosi et al. [10] exclusively reports that ternary nanofluids outperform single-particle systems. Farooq et al. [11] and Abbas et al. [12] studied mixed convection and cilia-driven flow in curved domains. Nabwey et al. [13] analyzed boundary layer flow over curved stretching surfaces. Heat diffusion, magnetic effects, and geometric variations in hybrid nanofluids are investigated by Jamshed et al. [14], Noreen [15], and Al Garalleh [16]. Use of machine learning is also proved to be an effective technique in predicting thermal performance of nanofluids. Patil et al. [17] examined mixed convection over rough cylinders under magnetic effects. Numerous studies have already addressed rheological and nanofluids flows under different physical configurations. The heat and mass transfer in non-Newtonian nanofluids within vertical enclosures have been examined by Hashim et al. [18] Rehman et al. [19] analyzed flow between intersecting planes with slip effects. Nadeem et al. [20] utilized a simple configuration to study the unsteady tangent hyperbolic hybrid nanofluid flow over exponentially stretching surfaces. Alqahtani et al. [21] studied ternary hybrid MHD nanofluids over wedge geometries. Alkaoud et al. [5] studied flow and temperature over stretching sheets in presence of magnetic field, porosity and Weissenberg number. Choudhary et al. [22] considered the porous and thermal aspects to study the boundary layer flow of tangent hyperbolic fluid in diverging channels. Unsteady flow assumption is more realistic than the steady flow in many systems, particularly to capture in turbulence, flow separation, and transient heat transfer. The work of Kebede et al. [23], Jamal et al. [24], and Reddy et al. [25] strongly argue that the unsteady parameters prominently affect velocity and thermal fields at higher Weissenberg numbers and under electromagnetic effects. Seeral works hae also been carried out with oscillatory flow and nonlinear stretching surfaces [22, 26]. Magnetohydrodynamics (MHD) is becoming popular among researchers as it finds important roles in astrophysics, nuclear cooling, and biomedical systems. MHD squeezing flow in porous media has been examined by Krishna and Chamkha [27]. The study exclusively reported that induced magnetic fields are negligible in water-based nanofluids due to low electrical conductivity. The Hall Effect is also weak under low velocity or weak magnetic field conditions. The nonlinear thermal radiation plays an important role in the heat transfer modelling of combustion, aerospace, geothermal, nuclear, and cooling systems [28–33]. The Stefan-Boltzmann law is nonlinear, the heat transfer behaviour is influenced when the thermal properties are temperature-dependent. Flow. Flow over stretching sheet is important in polymer extrusion, wire drawing and cooling systems. Interested readers can also be referred to other relevant studies on boundary layer flow, heat transfer and mass transfer over such surfaces for improvement in system design [34–38]. Al₂O₃–Cu–TiO₂ with ethylene glycol (EG) is selected based on an energetic consideration of the thermophysical properties of the three nanoparticles. Al₂O₃ is used for chemical stability and a moderate thermal conductivity (40 W/m·K); Cu for high electrical and thermal conductivity (400 W/m·K); and TiO₂ for oxidation resistance and photocatalytic activity. EG is chosen because of its wide applications in automotive cooling system and small heat exchanger, and low freezing point. This ternary combination of EG has been tested experimentally in the literature [39, 40, 64]. Moreover, some recent works have investigated the non-Newtonian nanofluids namely tangent-hyperbolic and Casson types with MHD conditions including radiation and viscous dissipation [41–45, 67, 68]. Several very recent studies further reinforce the growing interest in nanofluid- and hybrid-nanofluid-based heat transfer enhancement across allied configurations, including groundwater and aquifer thermal transport [46], interfacial and combined convective-radiative heat and mass transfer [47, 48], biological and biomechanical flow analogues [49, 66], thermal energy storage systems [50], convective transport in micro channels [51, 65], and hydrogen-based energy systems[52]. These studies collectively confirm the continued relevance of the present tangent-hyperbolic ternary hybrid nanofluid model and situate it within the current literature. Despite these advances, most studies consider only a limited set of physical effects. The combined influence of viscous dissipation, Joule heating, nonlinear thermal radiation, chemical reactions, thermophoresis, Brownian motion, Darcy–Forchheimer porous media, and variable thermal conductivity has not been examined for unsteady tangent hyperbolic ternary hybrid nanofluids. This gap limits the accurate prediction of transport behavior in complex systems. Table 1 shows a tabular comparison of the present study with the related literature in recent years. Table 1 shows that none of the above physical effect is considered in any previous studies on the flow of tangent-hyperbolic ternary hybrid nanofluid. No previous report has been found on the simultaneous treatment of all thirteen physical effects in an unsteady permeable stretching surface of Al₂O₃–Cu–TiO₂/EG ternary hybrid system. The main objective of the present study is to address this gap by developing a comprehensive model for unsteady tangent hyperbolic ternary hybrid nanofluid flow. The study considers a ternary nanofluid composed of Al₂O₃, Cu, and TiO₂ nanoparticles dispersed in ethylene glycol. The flow is modeled over a permeable stretching sheet, which is relevant to many industrial processes. The novelty of this work lies in the simultaneous inclusion of multiple interacting effects within a single framework. These include viscous dissipation, Joule heating, nonlinear thermal radiation, chemical reactions, thermophoresis, Brownian motion, Darcy–Forchheimer porous media, and variable thermal conductivity. Such a combined analysis has not been reported for this class of fluids as per the best of authors’ knowledge. Table 1Comparison of Present Study with the available literaturesStudy / ReferenceNanofluid SystemMHDUn- steadyTan- HypTernary HybridDa–F PorousNL Rad.Var. Cond.Visc. Diss.Joule Heat.Thermo/ Brown.Chem. React.Hussain et al. [53]Tangent-hyperbolic nanofluid / Water ✓ ✘ ✓ ✘ ✘ ✘ ✘ ✓ ✘ ✘ ✘ Nandi & Kumbhakar [54]Tangent-hyperbolic nanofluid / Water ✓ ✘ ✓ ✘ ✘ ✓ ✘ ✘ ✘ ✘ ✘ Jawad et al. [55]Nanofluid / Ethylene glycol ✓ ✘ ✘ ✘ ✓ ✓ ✘ ✘ ✓ ✘ ✓ Rafique et al. [39]Al₂O₃–Cu–TiO₂ / Ethylene glycol ✓ ✘ ✘ ✓ ✘ ✘ ✘ ✘ ✘ ✘ ✓ Manjunatha et al. [40]Ternary hybrid nanofluid / Water ✘ ✘ ✘ ✓ ✘ ✘ ✘ ✘ ✘ ✘ ✘ Waini et al. [56]Al₂O₃–Cu hybrid / Water ✘ ✓ ✘ ✘ ✘ ✘ ✘ ✘ ✘ ✘ ✘ Jamrus et al. [57]Hybrid nanofluid / Water ✓ ✘ ✘ ✘ ✓ ✘ ✓ ✘ ✘ ✘ ✘ Moltot et al. [6]Tangent-hyperbolic hybrid / Water ✓ ✓ ✓ ✘ ✓ ✘ ✘ ✘ ✓ ✘ ✘ Jan et al. [38]Hybrid nanofluid / Ethylene glycol ✓ ✘ ✘ ✘ ✘ ✓ ✘ ✓ ✘ ✘ ✘ Devi & Devi [58]Al₂O₃–Cu / Water ✘ ✘ ✘ ✘ ✘ ✘ ✘ ✘ ✘ ✘ ✘ Nabwey et al. [13]Ternary hybrid nanofluid / Water ✓ ✓ ✘ ✓ ✓ ✓ ✘ ✘ ✘ ✘ ✘ Present Study Al₂O₃–Cu–TiO₂ / Ethylene glycol (Ternary Hybrid) ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ ✓ Note: In Table 1, the feature is denoted as ✓ if it is part of the study and ✘ if it is not part of the study; The rheology is denoted as Tan-Hyp for Tangent-Hyperbolic, and as Darcy–F for Darcy-Forchheimer porous medium. Various thermal conductivity (Cond.) and viscosity values (Visc Diss., Joule Heat., Thermo/Brown, Chem.) have been considered for various components. React. = chemical reaction, heat is evolved/reacted with. Comparison of Present Study with the available literatures Note: In Table 1, the feature is denoted as ✓ if it is part of the study and ✘ if it is not part of the study; The rheology is denoted as Tan-Hyp for Tangent-Hyperbolic, and as Darcy–F for Darcy-Forchheimer porous medium. Various thermal conductivity (Cond.) and viscosity values (Visc Diss., Joule Heat., Thermo/Brown, Chem.) have been considered for various components. React. = chemical reaction, heat is evolved/reacted with. In this analysis the governing partial differential equations are transformed into ordinary differential equations using similarity transformations. The resulting system is solved numerically using the BVP5C solver in MATLAB. The effects of critical parameters on velocity, temperature, and concentration fields are examined. The impact of those parameters on skin friction, heat transfer, and mass transfer rates are also evaluated. The findings of this study will be useful for the design and optimization of thermal systems, biomedical devices, and energy applications. Methodology Problem description Figure 1 depicts the schematic of the physical flow model in Cartesian coordinate system. We investigate the transport of ternary hybrid nanofluid (THNF) over a permeable stretching sheet which is dictated by laminar boundary layer, two-dimensional, time-dependent, incompressible flow of an electrically conducting Tangent hyperbolic fluid. The THNF is driven by the speed of the stretching speed i.e. , where a is a positive constant, t is time, and c is the time-dependent parameter of this problem. The wall mass suction speed is considered as . The x-axis is oriented along the direction of the stretching sheet, whereas the y-axis is perpendicular to it. The velocity components are represented as u and v, corresponding to the x- and y-axes, respectively. The wall temperature, , ambient temperature,, well concentration, , and ambient concentration, of the THNF are considered constants. Schematic of the physical description of the problem The transport model consists of a set of governing equations including modified Navier-Stokes momentum, mass, and advective-diffusive energy transport equation containing thermophoresis, Brownian motion, viscous dissipation, chemical reactions, non-linear thermal radiation, heat generation, magnetic dissipation, and wall mass suction term. A constant magnetic field is applied perpendicular to the surface in the positive y-direction, the induced magnetic field is considered negligible and ignored. We also consider Darcy-Forchheimer flow of the THNF with variable thermal conductivity [59] in the present study assuming the porosity of the sheet remains constant. It is also assumed that thermophysical properties of the working fluid are constant and the system is in thermal equilibrium. The THNF is created through the dispersion of Al2O3, Cu, and TiO2 nanoparticles in C2H6O2 medium. Mathematical modelling The tangent hyperbolic model is widely employed to characterize the rheological behaviour of shear-thinning non-Newtonian fluids. In this model, the extra stress tensor is defined as: where the extra stress tensor E is given by Here, and denote the zero and infinite shear rate viscosities, respectively, is the material parameter, and is the power-law index. The shear rate is defined as: where is the second invariant of the strain rate tensor: Assuming and , the stress tensor simplifies to: The assumptions µ∞ = 0 and < 1 underlying Eq. (5) are standard for tangent-hyperbolic fluids at low-to-moderate shear rates, where the infinite-shear viscosity is negligible compared with the zero-shear viscosity and the product of the material time constant N and the shear rate remains below unity. For the parameter range explored in this study (We = 0.1–0.9, corresponding to well below unity at every η-station checked a posteriori from the converged f′′(η) profiles), both conditions remain valid throughout the boundary layer, so the simplified stress tensor of Eq. (5) is an appropriate and consistent approximation for the present flow regime. Assuming thermal equilibrium and no-slip conditions between the phases, the governing equations for the incompressible, laminar flow are: The Darcy–Forchheimer terms in Eq. (7) are for the inertia-dominated porous flow, which is extended from Darcy’s law, as detailed by Chamkha [60] and Jawad et al. [55]. The viscous dissipation and joule heating terms in Eq. (8) can be obtained by taking the inner product of the tangent-hyperbolic stress tensor (Eq. (5) with the strain rate tensor and the ohmic heating term . The assumption of thermal equilibrium and no slip is valid for φ ≤ 0.05. The boundary conditions are: To transform the governing partial differential equations into ordinary differential equations, the following similarity variables are employed: The velocity components are defined as and . The wall suction velocity is given in Eq. (11), with a single consistent expression used throughout the manuscript, where is the suction parameter. To verify the validity of the selected similarity transformations, Eq. (11) was substituted into the governing continuity, momentum, energy, and concentration equations. Using the unsteady stretching velocity the similarity variable , and the wall suction velocity , all explicit dependence on the variables (x) and (t) was eliminated after simplification, resulting in ordinary differential equations in terms of only. Further, the temperature-dependent thermal conductivity and nonlinear radiation terms were expressed in dimensionless form using which preserved the similarity structure without introducing additional independent variables. Therefore, the selected transformations are mathematically compatible with the assumptions of unsteady stretching, suction, variable thermal conductivity and nonlinear thermal radiation which are adopted in the present study. The similarity variables defined in Eq. (11) were substituted into the governing momentum, energy, and concentration equations, and the required derivatives were evaluated using the chain rule. After simplification and systematic cancellation of common factors, the explicit dependence on the spatial and temporal variables was eliminated, resulting in the transformation of the governing PDE system into the ordinary differential equations presented in Eqs. (12)–(14). Differentiating with respect to produces two terms, . The quadratic-gradient term is inserted in Eq. (13) and propagated to the first-order system, Eq. (22).The thermal-conductivity variation parameter is denoted Γ* while Γ is reserved exclusively for the tangent-hyperbolic parameter of Eqs. (2)-(5) and the Weissenberg number, Eq. (17). Subject to the transformed boundary conditions: The velocity ratio parameter characterizes the nature of the surface motion. Specifically, represents a stationary sheet, corresponds to a shrinking sheet, and indicates a stretching sheet. The thermophysical ratios of the ternary hybrid nanofluid are given by: The governing dimensionless parameters include the Weissenberg number We, magnetic parameter M, porosity parameter K, Forchheimer number Fr, Prandtl number Pr, nonlinear thermal radiation parameter R, temperature ratio parameter , heat generation/absorption parameter for heat generation and Q<0 for heat absorption ), Brownian motion parameter Nb, thermophoresis parameter Nt, Eckert number Ec, Schmidt number Sc, chemical reaction parameter Cr, and the unsteadiness parameter A. These parameters are defined as follows: Aside from the notation, it is important to note that the streamwise coordinate x in the definition of We and Fr in Eq. (17) is not an independent variable when these groups are introduced in Eqs. (12)-(14). Both parameters are evaluated at the fixed reference station used in the construction of the similarity variable η, in Eq. (11), (as in the case of tangent-hyperbolic and Darcy–Forchheimer stretching-sheet flows, e.g. Nadeem and Akram [4] and Chamkha [60] respectively), and once evaluated there, We and Fr are constants of the flow (functions of a, c, Γ, Cf, νf and the fixed reference x only) and hence have no residual (x, t) dependence in the ODE system. The thermophysical properties utilized for the base fluid (Ethylene Glycol) and the nanoparticles are summarized in Table 2. The correlations for estimating thermophysical properties are presented in Table 3. Table 2Thermophysical properties of base fluid and nanoparticles [61, 62].Physical PropertiesAl2O3CuTiO2C2H6O2 (EG)ρ (kg/m3)3970893342501115Cp (J/kg · K)765385686.22430k (W/m · K)404008.95380.253σ (S/m)3.5 × 1075.96 × 1071.0 × 10− 120.107 Thermophysical properties of base fluid and nanoparticles [61, 62]. Table 3Thermophysical properties correlations for ternary and hybrid nanofluid [39, 40, 58]PropertyTernary hybrid nanofluidHybrid nanofluid Dynamic viscosity Density Thermal conductivity where where where Heat capacity Electrical conductivity where where where Thermophysical properties correlations for ternary and hybrid nanofluid [39, 40, 58] where where where where where where Note that the values of the solid volume fractions of Al2O3, Cu and TiO2 (namely, those of , and , respectively) are constant parameters prescribed as input and are not dependent on the extent of reaction η. Equations (11) and (14) are called the dimensionless species field because it is a function of the dimensionless variable η. The transport of the dissolved chemical species accounted for by Brownian diffusion, thermophoresis (Eq. 9) and chemical reaction is an independent field variable (Eq. 11). In the present model the two are different physical entities (fixed number of nanoparticles versus diffusive concentration field) and hence do not double count the effects of nanoparticles. The physical quantities of interest are the skin friction coefficient and the local Nusselt Number and local Sherwood number : where , , and denote the wall shear stress, heat flux, and mass flux, respectively, given by: By applying the similarity transformations and substituting these gradients, the reduced expressions for the skin friction and heat transfer rate are obtained: Numerical method The transformed system of nonlinear ordinary differential equations given in Eqs. (11)–(13), together with the boundary conditions (14), is solved numerically using the bvp5c solver available in MATLAB. This solver is specifically developed for boundary value problems and is based on a collocation technique with adaptive mesh selection, which ensures reliable and accurate solutions for nonlinear and stiff systems. The bvp5c algorithm is particularly effective for problems involving multiple boundary conditions and coupled differential equations. Its adaptive mesh refinement capability improves solution accuracy while minimizing computational effort. In comparison with conventional finite difference schemes, this approach provides faster convergence and reduced dependency on discretization size, making it suitable for complex fluid flow and heat transfer analyses. To implement the numerical procedure, the higher-order governing equations are first rewritten as a system of first-order differential equations. This is accomplished by introducing the following transformations: Using these definitions, the original equations are converted into the following system: The corresponding boundary conditions are: where , , and are unknown initial slopes are determined during the solution process. The prime symbol denotes differentiation with respect to the similarity variable . The whole numerical procedure is presented as a flow chart in Fig. 2. Flow chart of the numerical procedure applied in the study Grid sensitivity A grid sensitivity analysis is carried out for the representative parameter set, A = 0.1, Nb = 0.5, Nt = 0.1 , , . Three different initial meshes used were 51, 101 and 201 grid points, and the convergence tolerance was kept constant at (10− 7). The differences between the results based on 101 and 201 initial mesh points for the skin-friction coefficient and local Nusselt number were determined as less than 0.005% and less than 0.01% respectively, which are found to be mesh independent to the desired degree of accuracy. Therefore, the initial mesh size of 101 was taken for all the following calculations (Table 4). The adaptive collocation solver bvp5c automatically adapted the mesh, which ended up at about 300–500 points, to meet the desired error tolerance (10− 7) for the solution process. The numerical computations are carried out using an initial mesh of 101 grid points with a step size of 0.1. The solver automatically refines the mesh to maintain accuracy, and a convergence tolerance of is imposed. This approach ensures stable and precise solutions for the coupled nonlinear system. Table 4Grid Independence StudyInitial Mesh PointsRex½ CfRex½ NuxRelative Change in Rex½ Cf (%)Relative Change in Rex½ Nux (%)511.1620820.132681------1011.1620670.1326740.00470.000792011.1620660.1326730.00080.00013 Grid Independence Study For reproducibility, the finite computational domain was truncated at ηmax = 10, a value increased from 6 to 10 in steps of 1 until the far-field boundary values f′, θ and Φ changed by less than 10− 5; the default parameter set used unless otherwise stated is that listed for Figure/Table baseline runs (A = 0.1, Nb = 0.5, Nt = 0.1, θw = 1.2, Q = 0.01, K = 0.1, λ = 1, Γ* = 0.3, We = 0.5, Cr = 0.1, R = 1.5, Ec = 0.2, Pr = 7, Sc = 2, n = 0.1, S = 0.1, Fr = 0.3, = 0.05, M = 0.2); the relative and absolute error tolerances of bvp5c (RelTol, AbsTol) were both set to 10−7, and the solver was run from the 101-point initial mesh reported above, with convergence confirmed by the mesh-independence check already described. Validation The reliability and accuracy of present numerical methodology was validated by comparing the present results with several well-established numerical and analytical solutions referred in open literature. Tables 5 and 6, and 7 present these comparisons for complex nanofluid configurations, demonstrating the consistency and precision of the present computational scheme across a wide range of physical conditions. Table 5Comparison between the numerical results and present results by evaluating 𝜃 (0) for (𝜑1=𝜑2=𝜑3 = 0) when 𝑀=𝑆=𝛽=0 and 𝜆=1PrJamrus et al.[55]Waini et al. [56]Gorla & Sidawi [57]Priyadlarshtni et al. [51]Present Study20.9113580.9113530.91140.91130.9113527716.131.7596851.759682––1.75968170271.8954031.89541.89541.89541.8954203.3539043.3539023.35393.35393.353901836 Comparison between the numerical results and present results by evaluating 𝜃 (0) for (𝜑1=𝜑2=𝜑3 = 0) when 𝑀=𝑆=𝛽=0 and 𝜆=1 Table 6andfor Al-water nanofluidwhen,(stretching case) and Pr=6.2 Waini et al. [56]Present Study 0.05-1.005381.62246-1.0053791.6224580.10-0.998771.49170-0.9987661.491690.15-0.981841.37543-0.9818511.375410.20-0.955921.27118-0.9559311.27116 andfor Al-water nanofluidwhen,(stretching case) and Pr=6.2 Table 7values for Al+Cu-water hybrid nanofluid when Pr=6.135,,(stretching case) and Rasool et al. [55]Devi & Devi [54]Present Results0.005-1.325862-1.327310-1.3270850.02-1.404648-1.409683-1.4094700.04-1.511257-1.520894-1.5206920.06-1.620177-1.634279-1.634083 values for Al+Cu-water hybrid nanofluid when Pr=6.135,,(stretching case) and Table 5 presents a comparison of the reduced Nusselt number for a pure fluid () under the limiting conditions and , across four Prandtl number values (). The present results are benchmarked against the numerical solutions of Jamrus et al. [57], Waini et al. [56], Gorla and Sidawi [41], and Priyadharshini et al. [62]. At Pr=2, the present value of agrees with the result of Jamrus et al. and Waini et al. [56] within a relative deviation of less than 0.0006% and 0.0001% respectively, while showing excellent agreement with the analytical result of Gorla and Sidawi ( 0.9114) to within 0.005%. At Pr=6.13, the result agrees very closely again with Jamrus et al. [63] and Waini et al. [56] respectively. At Pr=7, the value is in perfect agreement with all four references. At Pr=20, the output closely resemblances to the reported data in the references. The exceptionally low deviations (< 0.005%) across all Prandtl numbers with the references provides strong confidence in the accuracy of the present numerical scheme for the pure fluid case. Table 6 presents a comparison of the reduced skin friction coefficient and the reduced Nusselt number for the mono nanofluid Al₂O₃–water under the stretching case conditions, across four nanoparticle volume fractions , and 0.20. The present results are compared against the numerical solutions of Waini et al. [56]. For the skin friction coefficient, at , the present value of -1.005379 compares against the reference value of -1.00538, yielding a relative deviation of less than 0.0001%. At , the present result of -0.998766 agrees with -0.99877 to within 0.0004%. At and , the deviations remain below 0.001% and 0.001% respectively. For the Nusselt number, at , the present value of compares with 1.62246, a deviation of 0.001%. At , and 0.20, the deviations are 0.001%, 0.001%, and 0.002% respectively. Across all volume fractions and both output quantities, the maximum relative deviation between the present results and those of Waini et al. [56] does not exceed 0.002%, confirming the high fidelity of the present numerical approach for mono nanofluid configurations. It is further noted that, as the nanoparticle volume fraction increases from 0.05to 0.20, the magnitude of the skin friction coefficient decreases from 1.00538 to 0.95592 (a reduction of approximately 4.9%), while the Nusselt number decreases from to 1.27118 (a reduction of approximately 21.7%), trends that are correctly and consistently captured by the present numerical solution. Table 7 presents a comparison of the reduced skin friction coefficient for the hybrid nanofluid Al₂O₃+Cu+TiO2–water under the stretching case conditions, across four nanoparticle volume fractions , and 0.06. The present results are benchmarked against the numerical solutions of Rasool et al. [57] and the reference data of Devi and Devi [58]. At , the value of. agrees with Devi and Devi results within a relative deviation of 0.017%, and with Rasool et al. [57] to within 0.092%. At , the output is within 0.015% and 0.342% respectively compared to Devi and Devi [58] and Rasool et al. [57] At , the deviations become 0.013% and 0.622% respectively. At , the deviations are within 0.012% and 0.848% respectively. The present results consistently exhibit closer agreement with the Devi and Devi solutions with maximum deviations below 0.017% than with the Rasool et al. [57] results, which show somewhat larger deviations attributed to differences in the numerical methodology and solution scheme employed by Rasool et al. [57]. To the best of author’s knowledge, since ethylene glycol (EG)-based hybrid nanofluid data are not available in the existing literature for direct comparison, the validation in Table 5 is conducted using water as the base fluid. The close agreement among the simulated result and published outputs approves the accuracy of the underlying mathematical formulation and numerical implementation of the present model. Furthermore, the progressive increase in the magnitude of the skin friction coefficient from 1.327085 at to 1.634083 at with an increase of approximately 23.1% is consistently captured by all three solutions, confirming the physically correct response of the model to increasing nanoparticle loading. The comparison in Tables 5 and 6, and 7 among pure fluid, mono nanofluid, and hybrid nanofluid configurations demonstrates that the maximum relative deviations for all cases remain well within 1% with established solutions available the published literature. Therefore, the present study establishes numerical accuracy and physical consistency of the proposed computational framework prior to its application to the ternary hybrid nanofluid problem. Results and discussion In this section, we discuss the numerical solution for the time-dependent MHD hyperbolic tangent ternary hybrid nanofluid flowing over stretched surface. Simulations are carried out to determine temperature, velocity, concentration, local skin friction, Nusselt number, and Sherwood number under different physical parameter effects such as Weissenberg number (We), magnetic field parameter (M), spherical-shaped Al2O3,Cu, and TiO2 nanoparticle volume fraction (1, 2, 3) ,heat generation (Q), Forchheimer number (Fr), Prandtl number (Pr), nonlinear thermal radiation(R), unsteady parameter (A), temperature ratio parameter (θw), Eckert number (Ec), Schmidt number (Sc), chemical reaction (Cr), Porosity parameter (K), Brownian (Nb) and thermophoresis (Nt) diffusion parameter which are showcased via graphical and table illustrations. The choice of parametric values in this study is based on some physical relevance, the characteristics of materials properties and flow regimes that can potentially exist, the heat transfer conditions involved, and lessons learned from previous empirical studies. Mapping these values to particular physical situations increases the accuracy and relevance of the model, yielding better predictive results and ameliorating understanding with respect to the system behaviour. Effects of magnetic parameter and nanoparticle volume fraction Figures 3 and 4 show the changes of and for various values of the magnetic parameter M and nanoparticle volume fraction , under fixed conditions of S=2.2, , and . Based on Fig. 3, the absolute values of increase monotonically with both Mand . At M=0, the skin friction values are approximately -6.3, -8.0, and -9.5for , 0.15, and 0.2, respectively. These values increase in magnitude as M increases to 2,and are approximately -7.2, -9.3, and -11.8 which are increases of approximately 14.3%, 16.3%, and 24.2% respectively. This linear increase proves the synergistic effect of the magnetic field strength and the loading of nanoparticles to increase the wall shear stress. Comparison across the volume fractions at constant M=1, the same increase in in of about 27%, is obtained when changing from 0.1to 0.15 and, again, an additional increase in changing from to 0.2 yields an increase in the value 22%. This observation is well consistent with that made by Tiwari and Das [42] that reports an increase in nanoparticle volume fraction results in an increase in the skin friction coefficient. At physically, the more the nanoparticle is concentrated, the more the effective dynamic viscosity of the ternary hybrid nanofluid, and the more the resistive force at the wall surface. Moreover, the stronger M increases the Lorentz force, which is a result of interaction between the transverse magnetic field applied and the electrically conducting nanofluid. This retarding force of the body slows down the flow velocity normal to the direction of the magnetic field and the velocity gradient at the wall becomes steeper, thereby increasing the surface shear stress and hence the skin friction coefficient. The volume fractions utilized for the nanoparticles in this study are both experimentally stable and practically achievable as confirmed by Hussain et al. [39] and Manjunatha et al. [40]. The influence of magnetic parameter (M) and on The impact of M and on In contrast, Fig. 4 demonstrates that decreases gradually with increasing values of both Mand . At M=0, the Nusselt number values are approximately 13.62, 13.45, and 13.38 for , 0.15, and 0.2, respectively. As M increases to 2, these values decline to approximately 13.60, 13.42, and 13.32, corresponding to reductions of approximately 0.15%, 0.22%, and 0.45% respectively. These decreases are quite small in size, but the fact that they occur in all volume fractions is a clear indication of the suppressive nature of the two parameters, which are the magnetic parameter and the nanoparticle loading on convective heat transfer. The behaviour is in line with the findings of Rafique et al. [39], where the rise in the volume fraction of the nanoparticles is associated with the rise in the fluid viscosity that brings about increased resistance to fluid motion. It consequently leads to a reduction in the convective heat transfer capacity of the fluid, and a subsequent reduction in the local Nusselt number. Also, the magnetic field lowers fluid velocity, making convective mixing close to the thermal boundary layer ineffective and decreasing the rate of heat exchange between the surface and fluid even more. Effect of stretching parameter and unsteadiness parameter Figures 5 and 6 depict the influence of the stretching parameter and the unsteadiness parameter Aon and , respectively. As observed in Fig. 5, the skin friction coefficient exhibits a consistently declining trend which is, its magnitude increases in the negative direction as increases from 1 to 2 for all values of A considered. At , the skin friction values are approximately -2.5, -2.6, and -2.2 for A=-2, -3, and -4, respectively. As increases to 2, these values rise in magnitude to approximately -6.5, -6.2, and -5.8, representing dramatic increases in magnitude of approximately 160%, 138%, and 164% respectively. This sharp escalation in skin friction with the stretching parameter reflects the intensifying surface traction as the sheet is increasingly elongated. Moreover, for a fixed , a more negative unsteadiness parameter A results in comparatively less negative skin friction values across the range examined, suggesting that stronger unsteadiness moderates the wall shear stress to a degree. From a physical standpoint, as the stretching parameter increases, the surface area exposed to fluid interaction is extended. Both the enhancement of the wall-normal velocity gradient and the wall shear stress are physically consistent with the observed increase in , and the vertical shift of the streamwise momentum profile toward a larger boundary layer region is also consistent. The higher value of results in a higher value of wall velocity which directly causes higher wall normal velocity gradient, thus higher wall shear stress and skin friction. Influence of the stretching parameter and the unsteadiness parameter on Influence of the stretching parameter and the unsteadiness parameter Aon Figure 6 reveals a contrasting and physically complementary behaviour in the heat transfer rate, where increases consistently with the stretching parameter for all values of A. At , the Nusselt number values are approximately 14.32, 14.52, and 14.75 for A=-2, -3, and -4, respectively. As increases to 2, these values rise to approximately 14.65, 14.83, and 15.08, representing increases of approximately 2.3%, 2.1%, and 2.2% respectively. Furthermore, a more negative unsteadiness parameter Aconsistently yields a higher Nusselt number at any given , with the difference between A=-2 and being approximately 2.9% at and maintained at approximately 2.9% at , indicating that the enhancing effect of unsteadiness on heat transfer is sustained uniformly across the range of stretching. The physical reasoning underlying these trends is that the increase in surface area associated with larger reduces the streamwise fluid velocity near the surface, which in turn elevates the effective viscosity in the near-wall region and promotes greater thermal energy exchange between the sheet and the adjacent fluid. In addition, the stronger time-dependent acceleration associated with larger enhances fluid entrainment and thermal mixing within the boundary layer, collectively intensifying convective heat transfer and raising the local Nusselt number. Effect of suction parameter and unsteadiness parameter Figures 7 and 8 illustrate the combined effects of the suction parameter S and the unsteadiness parameter Aon and , respectively. As shown in Fig. 7, the absolute values of increase with higher suction strength S in the entire range of A examined. At A=0, the skin friction values are approximately -2.83, -3.03, and -3.28 for S=2.0, 2.2, and 2.4, respectively. When A decreases to -4, the magnitudes increase to approximately -2.07, -2.28, and -2.52. Comparing the suction values at A=-2, increasing S from 2.0- 2.2 results in an increase in of approximately 8.2%. A further increase from S=2.2-2.4 yields an additional rise of approximately 8.5%. This consistent increase indicates that suction exerts a proportionally uniform and significant influence on skin friction enhancement. The application of suction draws the fluid toward the stretching sheet surface causing the hydrodynamic boundary layer to thin considerably. A thinner velocity boundary layer results in a steeper velocity gradient at the wall which directly intensifies the surface shear stress and elevates the magnitude of the skin friction coefficient. Furthermore, at a fixed S, the magnitude of increases as decreases toward more negative values indicating that stronger flow unsteadiness amplifies the wall shear stress. At S=2.2, for instance, the increase in as A decreases from to -4 is approximately 24.7%. It reflects the substantial contribution of unsteadiness to momentum transport intensification near the wall. effects of the suction parameter Sand the unsteadiness parameter Aon effects of the suction parameter Sand the unsteadiness parameter Aon , Figure 8 demonstrates that the heat transfer rate also increases significantly with the suction parameter S for all values of A. At A=0, the Nusselt number values are approximately 12.75, 13.87, and for S=2.0, 2.2, and 2.4, respectively. As reduced to -4, these values increased to approximately 13.62, 14.85, and 15.93. When compared across suction values at a fixed A=-2, increasing S from 2.0-results in a Nusselt number enhancement of approximately 8.4%, while a further increase in - 2.4 produces an extra increase of approximately 7.9%. These quantitative increments highlight the strong and consistent boosting effect of suction on convective thermal transport. The underlying physical mechanism can be understood that the thermal boundary layer becomes thinner since the suction carries the fluid into closer proximity with the heated stretching sheet. A reduction in thermal boundary layer thickness develops a steeper temperature gradient at the wall. As a result, the convective heat flux from the sheet to the fluid enhances. Accordingly, the local Nusselt number rises with increasing suction strength, demonstrating that suction serves as an effective mechanism for simultaneously enhancing both momentum and thermal transport in ternary hybrid nanofluid flow. Additionally, for a fixed suction parameter S, the Nusselt number is observed to increase as A decreases toward more negative values. At S=2.4, for example, the Nusselt number increases by approximately 5.5% as A decreases from to -4, confirming that stronger unsteadiness in the flow enhances convective thermal transport by promoting greater fluid entrainment and intensifying thermal mixing within the boundary layer. These findings collectively demonstrate that suction and unsteadiness act as complementary and mutually reinforcing mechanisms that simultaneously augment both momentum and thermal transport characteristics in ternary hybrid nanofluid flow over a stretching surface. Effect of suction parameter on velocity, temperature, and concentration profiles Figures 9, and 10 illustrate the influence of the suction parameter S on the dimensionless velocity profile , temperature profile , and concentration profile , respectively, for S=2.0, 2.2, and 2.4. Velocity distribution for different values of 𝑆 Temperature distribution for different values of 𝑆 Figure 9 presents the velocity boundary layer profiles for the ternary hybrid nanofluid under varying suction strengths. All three profiles originate from at the wall, satisfying the no-slip stretching surface boundary condition, and decay monotonically toward zero as , confirming that the far-field boundary conditions are asymptotically satisfied. As the suction parameter S increases from 2.0 to 2.4, the velocity profile decays more rapidly with , indicating a progressive thinning of the hydrodynamic boundary layer. At , the velocity values are approximately 0.42, 0.38, and 0.33 for S=2.0, 2.2, and 2.4 respectively, reflecting reductions of approximately 9.5% and 13.2% relative to S=2.0. By , the velocity has fallen to approximately 0.10, 0.08, and 0.06 for the three suction values, representing decreases of approximately 20% and 40% relative to S=2.0. This accelerated decay of velocity with increasing Sis a direct physical consequence of the suction mechanism: as fluid is continuously withdrawn through the permeable stretching surface, the momentum of the fluid layer adjacent to the wall is reduced, pulling streamlines closer to the surface and compressing the extent of the velocity boundary layer. The boundary layer thickness therefore decreases monotonically with S, which is consistent with the well-established role of wall suction as a boundary layer control mechanism in convective flow problems. Figure 10 presents the temperature boundary layer profiles under the same suction conditions. Similar to the velocity profiles, all temperature profiles satisfy at the wall and decay asymptotically to zero far from the surface. As the suction parameter Sincreases from 2.0 to 2.4, the temperature profile decays considerably more steeply, indicating a marked thinning of the thermal boundary layer. At , the temperature values are approximately 0.50, 0.43, and 0.37 for S=2.0, 2.2, and 2.4 respectively, corresponding to reductions of approximately 14% and 26% relative to S=2.0. By , the temperature has fallen to approximately 0.10, 0.07, and 0.04, representing decreases of approximately 30% and 60% relative to S=2.0. The strong suppression of the thermal boundary layer with increasing Soccurs because suction brings cooler fluid from the free stream into closer contact with the heated sheet, simultaneously removing warm fluid from the boundary layer region and steepening the wall temperature gradient. This steeper gradient is directly responsible for the enhanced local Nusselt number observed in Fig. 7, establishing a clear and physically consistent link between the profile behavior and the heat transfer coefficient results. Furthermore, the thermal boundary layer is observed to be thinner than the velocity boundary layer for all values of S, which is consistent with the Prandtl number being greater than unity for ethylene glycol-based nanofluids. Figure 11 presents the nanoparticle concentration profiles under varying suction strengths. All profiles begin at at the wall and decay rapidly toward zero in the far field, satisfying the prescribed boundary conditions. The influence of suction on the concentration boundary layer follows the same trend observed for the velocity and temperature profiles: as Sincreases from 2.0 to 2.4, the concentration profile decays more rapidly with , signifying a reduction in the concentration boundary layer thickness. At , the concentration values are approximately 0.22, 0.18, and 0.14for S=2.0, 2.2, and 2.4 respectively, reflecting reductions of approximately 18.2% and 36.4% relative to S=2.0. This thinning of the concentration boundary layer with increasing suction strength can be physically attributed to the withdrawal of nanoparticle-laden fluid through the permeable surface, which reduces the diffusion depth of nanoparticles into the fluid domain and confines the concentration gradient to a thinner near-wall region. The consistency of this trend across all three boundary layer profiles such as velocity, temperature, and concentration confirm that suction acts as a globally effective boundary layer thinning mechanism for ternary hybrid nanofluid flow over a stretching surface, and all profiles are observed to satisfy the far-field boundary conditions asymptotically, validating the physical and mathematical consistency of the present solution. Velocity distribution for different Nanoparticles and concentrations Comparison of Nanofluid Types and Volume Fractions on Velocity and Temperature Profiles Figures 11 and 12 present a comparative analysis of the velocity profile and temperature profile for three distinct nanofluid configurations — namely, mono nanofluid (-EG), hybrid nanofluid (+Cu-EG), and ternary hybrid nanofluid (+Cu+-EG) — each evaluated at two nanoparticle volume fractions, (solid lines) and (dashed lines). Temperature distribution for different Nanoparticles and concentrations Figure 11 reveals that the velocity boundary layer thickness increases progressively as the nanofluid configuration advances from mono to hybrid to ternary hybrid and similarly increases with the nanoparticle volume fraction from to . At , the velocity values for the solid-line group () are approximately 0.42, 0.40, and 0.38 for -EG, +Cu-EG, and +Cu+-EG respectively, representing successive decreases of approximately 4.8% and 5.0%. For the dashed-line group at , the corresponding velocity values at the same are noticeably lower, approximately 0.35, 0.31, and 0.27, reflecting reductions of approximately 11.4% and 12.9% between successive nanofluid types. These results indicate that as additional nanoparticle species are incorporated into the base fluid and as the volume fraction is increased, the effective dynamic viscosity of the nanofluid rises substantially, retarding the fluid motion and causing the velocity to decay more slowly paradoxically resulting in a thicker velocity boundary layer due to the increased momentum diffusion. The ternary hybrid nanofluid at high volume fraction thus exhibits the most significantly extended velocity boundary layer among all configurations examined, which has direct implications for the wall shear stress and skin friction behavior discussed in the preceding sections. Figure 12 presents the corresponding temperature profiles and reveals a markedly contrasting and physically significant trend. As the nanofluid configuration advances from mono to hybrid to ternary hybrid, and as the volume fraction increases from to , the temperature profile decays considerably more rapidly with , indicating a pronounced thinning of the thermal boundary layer. At , the temperature values for the solid-line group () are approximately 0.55, 0.50, and 0.44 for -EG, +Cu-EG, and +Cu+-EG respectively, representing successive decreases of approximately 9.1% and 12.0%. For the dashed-line group at , the temperature values at the same location are dramatically lower, approximately 0.30, 0.20, and 0.12, reflecting decreases of approximately 33.3% and 40.0% between successive nanofluid types. This dramatic steepening of the temperature gradient with increasing nanoparticle complexity and volume fraction is a direct consequence of the substantially enhanced effective thermal conductivity of the ternary hybrid nanofluid. The incorporation of thermally superior nanoparticles particularly and in addition to into the ethylene glycol base fluid markedly improves the thermal diffusivity of the mixture, enabling more efficient conduction of heat away from the surface and into the fluid bulk. This results in a steeper near-wall temperature gradient, a thinner thermal boundary layer, and consequently a higher local Nusselt number, confirming that the ternary hybrid nanofluid configuration at elevated volume fractions offers the most superior thermal performance among all nanofluid types considered in the present study. All profiles in both Figs. 10 and 11 are observed to satisfy their respective far-field boundary conditions asymptotically, confirming the physical validity and numerical accuracy of the present boundary layer solutions. It is observed from Table 8 that +Cu+-EG gets always better performance than both the mono and hybrid nanofluids. In comparison to the mono nanofluid, the enhancement of the local Nusselt number is 50.22%, 111.05%, and 285.58% for = 0.05, 0.10, and 0.20 respectively, and over the hybrid nanofluid, the local Nusselt number is enhanced by 17.92% at the same = 0.05. The results quantitatively validate that the ternary hybrid nanofluids have a better thermal performance compared to other nanofluids and confirm that these nanofluids are suitable for high thermal efficiency applications. Table 8Percentage enhancement of ternary hybrid vs. mono nanofluids Comparison (%)= 0.05Mono → Hybrid21.69%27.38%= 0.05Hybrid → Ternary10.53%17.92%= 0.05Mono → Ternary34.51%50.22%= 0.10Mono → Ternary68.41%111.05%= 0.20Mono → Ternary151.13%285.58% Percentage enhancement of ternary hybrid vs. mono nanofluids The effects of the governing parameters on the skin friction coefficient , local Nusselt number , and local Sherwood number are presented in Tables 7, 8, 9, 10, 11, 12, 13, 14 and 15. The baseline parameter set is fixed, and each parameter is varied independently to quantify its relative influence through percentage variation. Table 9Effect of weissenberg number (We)We 0.10.5904331.7622461.0624271.50.5935551.6884651.0548203.00.5950711.5915291.0439553 Effect of weissenberg number (We) Table 10Effect of Magnetic Parameter (M) M 0.11.7422820.5914211.0604190.31.7039940.2880170.8483060.51.6711540.0464930.662184 Effect of Magnetic Parameter (M) Table 11Effect of Radiation Parameter (R) 0.01.6620620.6912771.0607530.71.9309860.3587571.0600061.52.2035560.0281531.060373 Effect of Radiation Parameter (R) Table 12Effect of eckert number (Ec) Ec 0.11.6732960.6525121.0606820.51.8088950.5325241.0602271.01.9663480.3936661.060035 Effect of eckert number (Ec) Table 13Effect of chemical reaction parameter (Cr) Cr 0.11.7422800.5914231.0604150.31.4435080.6347121.0316220.51.7098560.6343660.971354 Effect of chemical reaction parameter (Cr) Table 14Effect of stretching parameter () 1.01.7422820.5914231.0604191.11.9803310.4697801.1002971.32.4824720.1544731.182048 Effect of stretching parameter () Table 15Effect of thermophysical parameters (A) A 0.11.7422820.0198711.0918521.01.7422820.3923791.0671082.01.7422820.7884661.055788 Effect of thermophysical parameters (A) Influence of weissenberg number (We) In Table 9A change of Weissenberg number of 0.1 to 3.0 results in a marginal increase in skin friction (~ 0.79%), which means that fluid elasticity slightly increases the wall shear. Conversely, the Nusselt number reduces by an average of 9.69% and the Sherwood number by 1.74%. This behaviour proves that both thermal and mass transport are inhibited by viscoelastic effects, which increases the thickness of momentum and scalar boundary layers. Effect of magnetic parameter (M) The strongest effects on transport characteristics are found in the magnetic parameter. As presented in Table 10, under the increase in M increases between0.1 and 0.5, the skin friction reduces by an average of -4.08% and the Nusselt number reduces radically by-92.14. The number of Sherwood also declines significantly (~ 37.55%). Such a strong suppression is explained by the Lorentz force that slows down the movement of the fluids and reduces the effectiveness of convective transport, thus increasing both thermal and concentration boundary layers. Effect of radiation parameter (R) The effect of radiation parameter is presented in Table 11. The skin friction coefficient is increased significantly (up to 32.57%) due to thermal radiation, which means that the near-wall flow is accelerated. Nonetheless, the Nusselt number reduces drastically (~ 95.93%), which indicates that radiative heat flux spreads energy throughout the boundary layer, lowering the wall temperature gradient. The Sherwood number has been almost constant (< 0.1%), which supports the fact that radiation has very little effects on mass diffusion and has a minimal effect on the thermal transport. Effect of eckert number ( Ec) As presented in Table 12, Increasing the Eckert number to 1.0 increases the skin friction by about 17.51% as the viscous dissipation increases the fluid motion at the wall. On the other hand, the Nusselt number is lowered by an average of 39.66, with the internal heat generation lowering the temperature gradient. The Sherwood number is practically constant meaning that there is weak coupling between mass transfer and viscous heating. Effect of chemical reaction parameter (Cr) The parameter of the chemical reaction causes a non-monotonic change in the skin friction in which the first decrease (~ 17.16) is followed by partial recovery (see Table 13). There is a low increment in the Nusselt number (~ 7.26%), and a reduction in the Sherwood number (~ 8.40%). This indicates the uptake of nanoparticles in the boundary layer, which dilutes concentration fields but modestly increases thermal transport using coupled diffusion processes. Effect of stretching parameter (). In Table 14, dominating the flow behavior is the stretching parameter. The growth of by 0.3 leads to a considerable increase in skin friction (~ 42.48%) and Sherwood number (~ 11.47%), but the Nusselt number declines drastically (~ 73.87%). This implies that when the strength of stretching is increased, momentum and mass transport are intensified whereas thermal gradients are decreased because of increased convection. Effect of thermophysical parameters (A and n) In Tables 15 and 16, both parameters A and n do not influence the skin friction, and that proves their decoupling of the momentum equation. However, they greatly improve the heat transfer, with Nusselt number rising to about 286.9% when using A and 72.87% when using n. In the meantime, the Sherwood number decreases marginally which means that increased thermal diffusion is coupled by lower concentration gradients. Table 16Effect of thermophysical parameters (n) n 0.11.7422820.5304961.0640981.01.7422820.7580831.0503662.01.7422820.9170331.040817 Effect of thermophysical parameters (n) Effect of suction parameter (S) In Table 17, Suction Parameter S leads to a significant decrease in the Nusselt number (~ 76.38%) and an improvement in the Sherwood number (~ 4.35%). The skin friction is not changed. This action implies that suction fixes the boundary layer and enhances the transfer of the mass at the cost of thermal transfer. Table 17Effect of suction parameter (S) S 0.11.7422820.9572491.0378150.21.7422820.5914231.0604170.31.7422820.2260871.083005 Effect of suction parameter (S) Effect of schmidt number ( Sc) In Table 18, The Schmidt number has negligible influence on skin friction but strongly affects mass transfer. As increases, the Sherwood number rises significantly (~ 31.47%), while the Nusselt number decreases slightly (~ 4.06%). This reflects reduced mass diffusivity, which steepens concentration gradients near the wall. Table 18Effect of schmidt number ( Sc) Sc 0.11.7422820.5914231.0604190.31.7422820.5778651.2370850.51.7422820.5674111.394141 Effect of schmidt number ( Sc) A complete parametric sensitivity ranking for all 15 governing parameters is given in Table 19. Stretching parameter λ shows the most significant effect on skin friction (541%) and the Nusselt number (2691%), indicating that it plays a major role in the momentum and heat transfer. Unsteadiness parameter A leads to 480% fluctuations in and 222% in due to modulation of effective flow time-scale. The radiation parameter R and the temperature ratio are the most sensitive for thermal effects (Nux variations are more than 39%), and the mass transfer () is most sensitive to Sc (more than 157%) and Cr (more than 90%). The results are stable under slight variation of the parameters ( M, We, Nb, and , respectively), which causes small variations of their results (≤ 8%). This is an important point for experimental implementation in which it can be difficult to control the parameters precisely. Table 19Parametric sensitivity rankingParameterRangeMinMax%MinMax%MinMax%Stretching Param.0.5–2.00.4652.982541.1%0.0070.2072691.4%1.0171.65762.9%Power-law Index n0.1–2.00.1791.162550.7%—0.133—1.2771.49116.8%Unsteadiness A0.1–2.00.8421.16237.9%0.1330.770480.1%0.3971.277221.7%Temperature Ratio1.0–1.41.1621.162< 0.01%0.1080.15039.1%1.2771.277< 0.01%Radiation Param. R0.0–1.51.1621.1620.000%0.0010.133>>100%1.2771.2891.0%Suction Param. S0.1–0.31.0641.26518.9%0.1210.14519.4%1.1441.41723.9%Schmidt Number Sc1.0–4.01.1621.1620.000%0.1330.133< 0.02%0.7952.044157.2%Chemical Reaction Cr0.1–2.01.1621.1620.000%0.1330.133< 0.03%1.2772.42289.7%Eckert Number Ec0.1–0.41.1621.1620.000%0.1180.14019.0%1.2761.277< 0.1%Heat Generation Q0.01–0.51.1621.1620.000%neg0.133large1.2771.2850.7%Weissenberg No. We0.1–2.01.1311.1703.4%0.1320.1363.6%1.2681.2790.8%Thermophoresis Nt0.1–0.71.1621.1620.000%0.1310.1331.2%1.2771.4029.8%Brownian Motion Nb0.1–0.71.1621.1620.000%0.1320.1342.0%1.2711.3597.0%Variable Conduct.0.1–0.71.1621.1620.000%0.1280.1388.0%1.2771.277< 0.1%Magnetic Param. M0.1–0.51.1561.1802.0%0.1320.1330.7%1.2741.2770.2% Parametric sensitivity ranking Among all parameters, the magnetic parameter and radiation parameter R exert the strongest suppressive effects on heat transfer, while the stretching parameter and thermophysical parameters A and n significantly enhance thermal performance. Mass transfer is most sensitive to Sc, , and M, whereas skin friction is primarily governed by , R, and Ec. These findings highlight the competing roles of electromagnetic, thermal, and diffusive mechanisms in controlling transport processes in complex nanofluid systems. Conclusion This study investigated the unsteady MHD flow and heat transfer of a Tangent-Hyperbolic ternary hybrid nanofluid over a permeable stretching surface. Using the MATLAB BVP5C solver, we analysed the effects of various physical parameters. The following quantitative and qualitative conclusions are drawn: The velocity profile reduces as the magnetic parameter M increases, due to the retarding Lorentz force; increasing M from 0.1 to 0.5 raises the skin friction magnitude by approximately 14.3% to 24.2% (Sect. 3.1, Figs. 3 and 4). The Forchheimer number similarly increases inertial flow resistance and reduces the velocity profile, consistent with its role as a nonlinear drag term in the momentum equation (Eq. 7); a dedicated quantitative parametric sweep for is recommended as future work to establish its precise percentage influence on and . The velocity profile reduces as the magnetic parameter M increases, due to the retarding Lorentz force; increasing M from 0.1 to 0.5 raises the skin friction magnitude by approximately 14.3% to 24.2% (Sect. 3.1, Figs. 3 and 4). The Forchheimer number similarly increases inertial flow resistance and reduces the velocity profile, consistent with its role as a nonlinear drag term in the momentum equation (Eq. 7); a dedicated quantitative parametric sweep for is recommended as future work to establish its precise percentage influence on and . The stretching parameter significantly intensifies surface traction, increasing skin friction magnitude by up to 164% as moves from 1.0 to 1.3. The stretching parameter significantly intensifies surface traction, increasing skin friction magnitude by up to 164% as moves from 1.0 to 1.3. The ternary hybrid nanofluid exhibits superior cooling capabilities compared to mono and binary systems. Higher nanoparticle volume fractions and nonlinear radiation enhance the thermal gradient at the wall. Specifically, increasing the suction parameter S from 0.1 to 0.3 results in a 76.4% reduction in the local Nusselt number (Table 17), as the suction-driven thinning of the thermal boundary layer steepens the wall temperature gradient in a manner that decreases the dimensionless Rex−1/2Nux. Suction more directly enhances skin friction and mass transfer (Shx increases by 4.35% over the same range) The variable thermal conductivity also acts as a key driving force for heat transfer enhancement. The ternary hybrid nanofluid exhibits superior cooling capabilities compared to mono and binary systems. Higher nanoparticle volume fractions and nonlinear radiation enhance the thermal gradient at the wall. Specifically, increasing the suction parameter S from 0.1 to 0.3 results in a 76.4% reduction in the local Nusselt number (Table 17), as the suction-driven thinning of the thermal boundary layer steepens the wall temperature gradient in a manner that decreases the dimensionless Rex−1/2Nux. Suction more directly enhances skin friction and mass transfer (Shx increases by 4.35% over the same range) The variable thermal conductivity also acts as a key driving force for heat transfer enhancement. Wall suction serves as an effective mechanism for boundary layer control. Increasing suction from 0.1 to 0.3 reduces the velocity atby nearly 40%, causing both the hydrodynamic and thermal boundary layers to thin. Because the Prandtl number for ethylene glycol is greater than unity (Pr > 1), the thermal boundary layer remains consistently thinner than the momentum boundary layer. Wall suction serves as an effective mechanism for boundary layer control. Increasing suction from 0.1 to 0.3 reduces the velocity atby nearly 40%, causing both the hydrodynamic and thermal boundary layers to thin. Because the Prandtl number for ethylene glycol is greater than unity (Pr > 1), the thermal boundary layer remains consistently thinner than the momentum boundary layer. The nanoparticle concentration decreases with higher Schmidt numbers and chemical reaction rates . However, thermophoresis tends to push particles away from the hot surface, increasing the concentration in the fluid bulk. The nanoparticle concentration decreases with higher Schmidt numbers and chemical reaction rates . However, thermophoresis tends to push particles away from the hot surface, increasing the concentration in the fluid bulk. The highest suppression was given by the parameters M and R with the values of 92.14% and 95.93% respectively in heat transfer () among all the parameters. The highest Nux enhancements (3868% and 72.87%) were observed with parameters A and n. Mass transfer is most sensitive to Sc (31.47%) and (11.47%), while skin friction () is primarily governed by (42.48%), R (32.57%), and Ec (17.51%). The highest suppression was given by the parameters M and R with the values of 92.14% and 95.93% respectively in heat transfer () among all the parameters. The highest Nux enhancements (3868% and 72.87%) were observed with parameters A and n. Mass transfer is most sensitive to Sc (31.47%) and (11.47%), while skin friction () is primarily governed by (42.48%), R (32.57%), and Ec (17.51%). This study provides valuable theoretical insights but has some limitations. Simplified assumptions, such as treating the nanofluid as a single-phase fluid and neglecting higher-order effects, may not fully capture real-world complexities. The BVP5C numerical method, while effective, has constraints in handling highly nonlinear systems. The findings are based on a specific parameter range, and extrapolation beyond this range requires further validation. Additionally, the assumption of uniform nanoparticle dispersion overlooks possible aggregation or sedimentation effects. To enhance validity and generalizability, future research should incorporate experimental validation and more advanced modeling approaches. Acknowledgements The authors gratefully acknowledge the support and facilities provided by their institution, which enabled the successful conduct of this research work. Author contributions ZG.DQ and M.Y wrote the main manuscript text, YN.L and TZ.H prepared figures 1-5, C.X Analyze the data in this article, GY editing and final approval. Funding No funding was received for this study. Data availability The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request. Code availability The computational code developed and used in this study is proprietary to the authors and is therefore not publicly available. Declarations Ethics approval. Not applicable. Consent to participate. Not applicable. Consent to publish. Not applicable. Competing interests The authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Data Availability Statement The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request. The computational code developed and used in this study is proprietary to the authors and is therefore not publicly available.