Model construction and knitting damage mechanism of extracorporeal membrane oxygenation membrane fabric

Abstract

Extracorporeal Membrane Oxygenation (ECMO) relies critically on the gas exchange performance of poly-4-methyl-1-pentene (PMP) membrane fabrics, which can be significantly affected by mechanical damage during the knitting process. However, the damage mechanisms induced by yarn tension during membrane fabric formation remain insufficiently understood. In this study, a combined geometrical modeling, finite element simulation, and experimental approach was employed to systematically investigate the mechanical deformation and functional degradation of PMP membranes during warp-knitting. A three-dimensional stitch geometry model of ECMO membrane fabrics was constructed and integrated into a finite element framework to simulate yarn tightening and contact-induced compression under different yarn tensions. The simulation results revealed a distinct elastic–plastic transition in the PMP membrane when yarn tension exceeded approximately 0.2 N, characterized by pronounced logarithmic and equivalent plastic strain localization. Experimental validation through outer diameter measurements, cross-sectional SEM observation, and porosity analysis demonstrated strong agreement with the simulation predictions. Below 0.2 N, membrane deformation was predominantly elastic and reversible, whereas higher tensions led to irreversible plastic flow, lumen collapse, wall folding, and pore closure. Quantitative pore structure analysis showed that exceeding the critical tension resulted in a significant reduction in open porosity and a marked increase in closed pore and wall volume fraction, directly impairing gas exchange capability. These results collectively establish 0.2 N as a critical yarn tension threshold for maintaining the structural integrity and functional performance of PMP-based ECMO membrane fabrics. This work elucidates the knitting-induced damage mechanism of ECMO membrane fabrics and provides a quantitative theoretical basis for tension control and low-damage manufacturing in large-scale industrial production.

Keywords

ECMO technology PMP membrane fabrics yarn tension gas exchange properties finite element simulation

Introduction

Extracorporeal Membrane Oxygenation (ECMO) is widely employed in the treatment of critically ill cardiopulmonary patients. The procedure involves circulating the patient’s blood outside the body, passing it through a membrane fabric for gas exchange to deliver oxygen and remove carbon dioxide.^1–4 Membrane fabrics serve as the core component,^5,6 and their performance directly influences the effectiveness of the treatment. As technological applications continue to expand,^7,8 optimizing the design and fabrication processes of membrane fabrics^9,10 to enhance gas exchange efficiency and biocompatibility^11,12 has become a central focus of current research.

While several studies have explored the impact of this technological process on the properties of membrane fabrics,^13,14 a comprehensive analysis of their damage mechanisms remains lacking. The fabrication of ECMO membrane fabrics is based on weft-insertion warp-knitting technology, a specialized branch of warp knitting, and understanding the specifics of how this preparation influences the membrane’s structure,¹⁵ mechanical properties,¹⁶ gas exchange efficiency,^17,18 and clinical stability is crucial.^19,20

Current research on ECMO membrane fabrics primarily focuses on the optimization of membrane material properties and surface modification.^21–24 Researchers have investigated how different materials affect gas exchange performance and how variations in factors such as membrane porosity and surface roughness influence gas permeability.^25–27 However, few studies have examined the specific effects of stitch yarn formation on the evolution of characteristics and structure during membrane fabric manufacture.^28–31 Some studies have shown^32,33 that yarn tension can reduce the porosity of membrane fabrics, thereby affecting gas exchange performance, but the damage mechanisms and specific outcomes remain poorly understood. Poly-4-methyl-1-pentene (PMP) represents the current gold standard material for ECMO oxygenators. Structurally, PMP hollow fibers are characterized by an asymmetric architecture comprising a dense, non-porous outer skin and a microporous support wall. This unique configuration allows for gas exchange via a solution-diffusion mechanism through the skin layer-effectively preventing plasma leakage inherent to conventional microporous polypropylene membranes—while maintaining high gas permeability. Consequently, PMP membranes are recognized as the most effective oxygenating membrane material in ECMO technology.^34–37 However, comparative evaluations of the performance of PMP membranes and membrane fabrics are often conducted separately,^38,39 failing to compare the performance of PMP membrane fabrics before and after their formation. Accordingly, this study integrates finite element simulation with experimental validation to quantitatively evaluate the influence of yarn tension on the structural stability and functional performance of PMP membrane fabrics.

To simulate the ECMO membrane fabric preparation process, a 3D model of the PMP membrane fabric is first developed, capturing its physical structure and dimensions. Subsequently, a finite element simulation model is constructed to integrate the forces involved during the preparation process. This model is used to calculate the morphological responses of the PMP membrane under varying yarn forces. The simulation also enables an in-depth examination of how different yarn forces affect the membrane’s structure. Finally, the influence of various knitting tensions on the damage mechanisms of ECMO membrane fabrics is systematically investigated through a combination of experimental analysis and computational modeling. The results of this study will provide essential theoretical insights into the optimal design and industrial production of ECMO membrane fabrics, while also offering a valuable reference for the future development and optimization of membrane materials by elucidating the role of yarn tension in shaping membrane fabric performance.

Materials and experimental methods

Materials and fabric preparation

In this study, 50D/36F low-elastic polyester filament yarn and high-performance PMP membranes were selected as the primary materials to investigate the loop morphology within the membrane fabric structures. The membrane fabric was manufactured using a TM-WEFT (E24) weft-insertion warp knitting machine, adhering to protocols established in prior studies.^40,41 To mitigate the influence of processing variability, the fabrication parameters were strictly controlled as follows: a machine speed of 50 r·min⁻¹, a threading arrangement of 1-full/10-empty, a lapping movement of 1–0//, a warp run-in of 1000 mm·rack⁻¹, and a threading density of 10 filaments·cm⁻¹. The resulting membrane fabric configuration is illustrated in Figure 1. The key geometric parameters of the stitch loops were quantified experimentally. Furthermore, a three-dimensional simulation model was developed to precisely characterize the morphology and structural features of both the PMP membrane and the constituent loops.

Figure 1.

ECMO membrane fabric with stitch size structure: (a) PMP membrane, (b) ECMO membrane fabric, (c) stitch size measurement, and (d) tension data acquisition.

During the weaving process, the warp tension data were collected in real time for subsequent simulation analysis. An online tension measurement platform was established, as shown in Figure 1(d). In the figure, E denotes the guiding and reversing position illustrated in Figure 3; P represents the weft-laying carriage; and Q indicates the braiding mechanism. Component a is the tension sensor, b is the LMS data acquisition unit, and c is the computer equipped with LMS software. The tested warp yarn was 50D polyester filament. The tension sensor was installed at position P, close to the weft-laying carriage, to acquire the yarn tension signal in real time. The tension signal was transmitted to an industrial computer through the data acquisition front-end and analyzed in both the time and frequency domains using LMS Test.Lab software.

Geometrical modeling and characterization

Stitch geometry measurement

The ECMO membrane fabric utilizes a warp-knitting technology characterized by the use of pillar stitches,^42–44 a structure primarily composed of stitch arcs and pillars. To ensure statistical reliability and minimize random errors, each principal dimensional parameter was measured at 10 randomly selected positions, after which the mean and standard deviation were calculated.

The principal measuring parameters are delineated as follows:

(a) Vertical height of the stitch (H_v): the distance from the apex of the stitch arc to the axial orientation of the stitch pillar.

(b) Actual height of the stitch arc (H_a): the vertical distance from the peak of the stitch arc to the base of the stitch.

(d) Height of the stitch arc (H_l): the maximum predicted length of the loop arc in the vertical axial orientation.

(e) Width of stitch (W_c): the maximum lateral extent of the stitch measured from the central axis of the yarn.

The dimensional data for each component of the stitch for this organization are presented in Table 1; the statistics reveal the dimensional variations of the stitch at various locations. For the membrane fabric organization, the average ratio of the vertical height (H_v) to the actual height (H_a) is 0.988, suggesting that the two measurements are nearly equivalent.

Table 1.

Stitch size.

Organizational type	Parameter	$H_{v}$ (µm)	$H_{a}$ (µm)	$H_{p}$ (µm)	$H_{l}$ (µm)	$W_{c}$ (µm)
Pillar stitch	Mean value	593.15	600.35	471.55	128.45	243.25
Pillar stitch	Standard deviation	4.55	6.35	7.25	6.34	8.35

Stitch model construction and 3D modeling of membrane fabrics

A 3D geometric model of the ECMO membrane fabric stitches was constructed using the “Textile AI Design System iTDS 3.0,”^45,46 a proprietary textile simulation software developed by our research group at the Engineering Research Center of Knitting Technology, Ministry of Education, Jiangnan University, as seen in Figure 2(a).The geometric model is defined by eight control points (characteristic points): six points govern the primary structure of the stitch loop, while the remaining two determine the extension lines.

Figure 2.

Stitch mesh model, type-value point distribution and 3D simulation: (a) 3D geometric model of the stitch, (b) stitch mesh model with type-value point distribution, and (c) 3D simulation results of the ECMO membrane fabric.

The definitions of these control points are as follows:

The stitch pillars are represented by straight lines connecting positions P₂–P₃ and P₆–P₇. The height of the stitch pillar, H_p, is determined by equation (1).

The root of the stitch pillar is positioned at the yarn center, with an offset determined by half the yarn diameter, as described in equation (2).

The stitch width, w is defined as the maximum lateral span at the base of the stitch arc, calculated via equation (3).

The stitch arc region is divided into three equal segments based on the stitch width. Points are assigned at 60° intervals, with coordinates defined by equations (4) and (5).

\begin{matrix} H_{p} = | P_{3} . y | = | P_{6} . y | = α H_{v} \end{matrix}

(1)

d_{w} = | P_{2} . x | = | P_{7} . x | = \frac{d}{2}

(2)

\begin{matrix} w = | P_{3} P_{6} . x | = β H_{v} \end{matrix}

(3)

\begin{matrix} | P_{4} . y | = | P_{5} . y | = H_{v} \end{matrix}

(4)

\begin{matrix} | P_{4} . x | = | P_{5} . x | = γ H_{v} \end{matrix}

(5)

where the coefficients α, β, and γ are determined from the data in Table 1 and take the following values: α = 0.79 H_v, β = 0.44 H_v, and γ = 0.14 H_v.

A planar assembly model was established based on the geometric datum of the stitch bottom and the stitch arc center. To capture the spatial distribution, a 2 × 2 grid model was constructed, as shown in Figure 2(b). The model defines the origin coordinate O(x₀, y₀, z₀) as the center of the current stitch pillar, and O₁(x₁, y₁, z₁) as the center of the adjacent transverse stitch pillar. The vertical distance between these points determines the course height (stitch pillar height offset), b = y₁ − y₀.

The full three-dimensional structural simulation was generated using the iTDS 3.0 system by applying the control point laws and the measured fabric parameters, as shown in Figure 2(c). A comparison with the physical samples confirms that the simulated model accurately replicates the appearance and dimensional ratios of the experimental fabric, validating the model’s reliability for subsequent mechanical analysis.

Simulation model and mechanical parameterization

Determination of material prameters

Accurate definition of material properties—specifically density, Young’s modulus, Poisson’s ratio, and yield strength—is a prerequisite for the high-fidelity finite element simulation of ECMO membrane fabrics. Unidirectional tensile tests were performed to characterize the mechanical constitutive behavior of both the PMP membrane and the polyester yarn, using the experimental setup illustrated in Figure 3(a).

Figure 3.

Simulation calculation process: (a) Material tensile strength test; (b) Model construction; (c) Boundary conditions and mesh.

Due to the initial preload applied during the tensile testing of PMP membranes, the stress-strain curve did not originate from absolute zero. Consequently, Young’s modulus (E) was determined by linearly fitting the data within the 0%–1.25% elongation range. The calculation procedure is as follows:

(a) Initial Cross-Sectional Area (A₀): Assuming a hollow cylindrical geometry for the PMP fiber, A₀ is calculated using the outer diameter (D = 0.38 mm) and inner diameter (d = 0.20 mm):

\begin{matrix} \begin{array}{l} A_{0} = \frac{π}{4} (D^{2} - d^{2}) \\ = \frac{π}{4} ({0.38}^{2} - {0.2}^{2}) \\ = 8.2 * 10^{- 2} m m^{2} \end{array} \end{matrix}

(6)

(b) Stress and Modulus Calculation: The engineering stress (σ) is derived from the applied force (F) and A₀. Note that force measured in centinewtons (cN) is converted to Newtons (N) for consistency.

\begin{matrix} σ = \frac{F (N)}{A_{0} (m m^{2})} = \frac{F (c N) \times 0.01}{0.082} (M P a) \end{matrix}

(7)

Young’s modulus is then calculated from the slope of the stress-strain curve within the linear elastic region (strain ε from 0 to 0.0125):

\begin{matrix} E = \frac{σ_{i -} σ_{0}}{ε} = \frac{σ_{i -} σ_{0}}{0.0125} \end{matrix}

(8)

(c) Material Constants: Based on these calculations, the average Young’s modulus of the PMP membrane was determined to be E_avg ≈ 147 MPa. The yield strength was approximated by linear interpolation at a strain offset of ε = 0.2% (0.002 strain), resulting in σ_avg ≈ 2.28 MPa.

The mechanical parameters for the polyester yarn were characterized using an identical methodology. To facilitate the finite element simulation, the multifilament yarn was treated as a homogenized continuum. The density and Poisson’s ratio were adopted from standard literature values for polyester and PMP materials.²⁰ The final material properties used in the simulation are summarized in Table 2.

Table 2.

Material parameters of polyester yarn and PMP membrane.

Material name	Density (kg/m³)	Young’s modulus (MPa)	Poisson’s ratio	Yield strength (MPa)
PMP membrane	880	147.0	0.35	2.28
Polyester yarn	1380	2700	0.32	80

Finite element model setup

To investigate the mechanical impact of yarn tension (F) on PMP membranes during the fabrication process, a finite element model was established, as illustrated in Figure 3(b). The model simulates the loop formation phase, where the knitting needle descends, applying tension F to tighten the loop. This action causes the yarn to slip and constrict, exerting compressive contact pressure on the PMP membrane surface and inducing deformation.

In this work, simulation calculations were performed using the finite element program Abaqus CAE-2021 edition. To increase the calculation accuracy of the process of the action of the PMP film on the yarn, only one longitudinal stitch of the knitted yarn was chosen as the research item.

The ECMO membrane fabric adopts a pillar stitches structure, in which the two components are in direct surface contact in the actual textile configuration. In the present numerical simulation, a surface-to-surface contact formulation was employed to characterize the interaction between these components. The PMP weft yarns were defined as the contact surface, whereas the polyester warp yarns were specified as the target surface.

Tension amplitude was measured under various knitting settings ranging from 0.1 0.45 N using an online tension sensor installed on the knitting machine (as described in Section 2.1). Fixed constraints were imposed in the finite element model at points e₁, e₂, and e₃, while the yarn end force was applied at position e₄. To ensure both computational efficiency and accuracy, a polyhedral mesh was used for both the PMP membrane and the yarn, as shown in Figure 3(c). The meshing results show consistent and complete mesh quality for both components.

The finite element simulation was performed using Abaqus/Explicit (dynamics module) due to the highly nonlinear nature of the problem, which involves material plasticity, large deformation of the PMP membrane, and surface-to-surface contact between the yarn and membrane. Although the yarn tension F was applied as a constant load within each simulation case (0.1–0.45 N), the deformation process was treated as a quasi-static analysis using the explicit solver. This approach was necessary because the implicit solver encountered severe convergence difficulties caused by the complex contact interactions and the abrupt onset of plastic instability when the tension exceeded the critical threshold.

To ensure quasi-static conditions and minimize inertial effects, the load was ramped smoothly using a smooth step amplitude curve over a total simulation duration of T = 0.02 s. This duration was selected based on a preliminary eigenfrequency analysis of the PMP membrane, ensuring that the applied loading rate was sufficiently slow to maintain the ratio of kinetic energy to internal energy below 5% throughout the analysis, as recommended for quasi-static explicit simulations. A fixed time increment of Δt = 1 × 10⁻⁶ s was used, which satisfies the Courant stability condition for the smallest element size in the mesh. The explicit solver was chosen over an implicit static solver because the latter failed to converge at higher tension levels (F > 0.2 N) due to severe plastic localization and contact instability.

Experimental characterization methods

To validate the simulation predictions and quantitatively assess the impact of knitting tension on the PMP membrane, the fabricated membrane fabric samples (produced under the varying yarn tensions F specified in Section 2.1) were subjected to the following characterization protocols.

Measurement of PMP membrane outer diameter

The variation in the outer diameter of the PMP membrane at the knitting contact point was selected as the primary indicator of macroscopic deformation. Prior to measurement, the knitting stitches were carefully removed to expose the deformed membrane surface. The outer diameter was quantified using a CU fiber fineness analyzer, see Figure 5(a), following the standard procedure GB/T 38902-2020. For each tension condition (F), 10 independent measurements were conducted at random locations. The mean diameter and the coefficient of variation (CV) were calculated to evaluate both the magnitude of deformation and the uniformity of the structural response.

Cross-sectional morphology of knitting positions

To investigate microstructural deformation mechanisms, the cross-sectional morphology of the membrane at the knitting position was examined using Scanning Electron Microscopy (SEM). The PMP membrane was obtained after weaving by non-destructively removing the yarns from the membrane fabric under different warp tensions. The cross-section at the knitting position was stabilized using resin and subsequently obtained via a cutting method. The pore sizes were then measured using SEM. Analysis of the resulting micrographs enabled qualitative and quantitative assessment of geometric distortions—specifically the transition from circular to elliptical profiles—and identification of critical defects, such as plastic flow, wall folding, or pre-cracking, particularly in samples subjected to high yarn tension, see Figure 6.

Porosity and pore structure analysis

The pore structure of the deformed PMP membranes was characterized using a BSD-TD porosity analyzer (Beishide Instrument, Beijing, China), as shown in Figure 7(a). This instrument operates on the principle of gas expansion displacement based on Archimedes’ principle and the Boyle–Mariotte law (PV = nRT).

The measurement procedure is as follows: First, the sample is placed in a sealed test chamber of known volume. An inert gas (N₂) is introduced into a reference chamber at a precisely controlled pressure. The gas is then allowed to expand into the sample chamber, and the equilibrium pressure is recorded. By comparing the pressure changes before and after gas expansion, the true volume of the solid sample (including closed pores) is calculated. The true density is then obtained as the ratio of sample mass to this true volume.

Open Porosity (OP): Defined as the percentage of interconnected pore volume accessible for fluid transport, which serves as a direct proxy for gas permeability.

Closed Pore and Wall Volume Percentage (PC): Defined as the combined volume fraction of sealed pores and solid polymer walls. An increase in PC indicates the collapse of effective gas pathways and a reduction in the functional surface area.

This gas displacement method was specifically selected for this study because it offers distinct advantages for porous membrane materials: (1) it uses inert gas that penetrates microporous structures without causing damage, (2) it avoids the dissolution issues associated with liquid immersion methods, and (3) it complies with the ISO 12154 and GB/T 10799-2008 standards for density and porosity determination of porous materials.

Results and discussion

This section presents a systematic investigation into the impact of yarn tension on ECMO membrane fabrics, progressing from the fundamental mechanical response of the PMP membrane to the consequent degradation of its functional pore structure. The findings from finite element simulation and experimental characterization are integrated to validate the model and elucidate a comprehensive damage mechanism.

Mechanical response and strain analysis via simulation

The finite element simulation provided a fundamental understanding of the stress distribution and deformation mechanisms of the PMP membrane under knitting loads. The deformation behavior of the longitudinal cross-section at the knitting position was quantified using Logarithmic Strain (LE) and Equivalent Plastic Strain (PE), as shown in Figure 4.

Figure 4.

Deformation results of PMP membrane with different forces F: (a) SEM image of PMP membrane; (b) statistical analysis of LE and PE results; € LE results of 0.1~0.25N; (d) PE results of 0.1~0.25N; (e) LE results of 0.3~0.45N; (f) PE resultsof 0.3-0.45N.

Figure 4(a) illustrates the progressive distortion of the PMP membrane cross-section. In the unloaded state, both the inner and outer apertures retain a circular geometry. As the yarn tension (F) increases, the membrane undergoes distinct deformation stages:

Elastic Region (F ⩽ 0.15 N): The LE and PE values remain low (LE ≈ 0.072, PE ≈ 0.067), indicating that deformation is primarily elastic and reversible.

Elastic-to-Plastic Transition (0.15 N < F ⩽ 0.2 N): A critical transition occurs within this force range. The outer diameter begins to contract, and as the tension increases, localized yielding initiates in the material, marking the onset of plastic deformation.

Plastic Failure Region (F > 0.2 N): Crucially, a failure region is identified when tension exceeds 0.2 N. Under these conditions, significant plastic flow occurs. At F = 0.3 N, the inner aperture—vital for gas flow—exhibits noticeable deformation. At extreme tension (F = 0.45 N), LE and PE values spike to 1.43 and 1.35, respectively. This indicates severe, irreversible collapse of the membrane wall structure, suggesting that the material at the knitting point is approaching mechanical failure.

These simulation results establish a theoretical critical threshold of 0.2 N, beyond which the risk of permanent structural damage increases exponentially.

Validation of outer diameter changes

To validate the simulation model, the macroscopic deformation of the PMP membrane was experimentally characterized by measuring the outer diameter at the knitting position, as shown in Figure 5. The experimental results demonstrate a high degree of correlation with the simulation predictions. The unknitted PMP membrane exhibited a consistent average outer diameter of 395 μm. As shown in Figure 5(a), the diameter decreases monotonically with increasing yarn tension.

Figure 5.

PMP membrane diameter tests for knitting experiments: (a) results of PMP membrane outer diameter changes under different force values F, (b) unknitted PMP membrane diameter, and (c) PMP membrane diameter at knitting position.

At tensions below 0.2 N, the reduction in diameter is linear and minimal, with a low coefficient of variation (CV < 5%), confirming the stability of the membrane in the elastic deformation phase. However, instability arises at high tension. Once tension exceeds the 0.2 N threshold identified in the simulation, the rate of diameter reduction accelerates significantly. Notably, the deformation becomes highly non-uniform, evidenced by a sharp rise in the coefficient of variation (CV reaching 22.35% at 0.35 N and 35.72% at 0.45 N). This instability in diameter measurement strongly supports the simulation finding that high tension induces plastic flow, leading to unpredictable and irregular structural collapse rather than uniform elastic compression.

Microstructural evolution and damage mechanisms

The macroscopic diameter reduction is driven by microstructural evolution, as revealed by the cross-sectional SEM morphology in Figure 6. The SEM imagery provides visual confirmation of the damage mechanisms inferred from the LE/PE simulation data.

Figure 6.

Cross-sectional SEM results of knitting position.

When F ⩽ 0.15 N, morphological stability is maintained; the cross-section remains quasi-circular, the reduction in the outer diameter is slight (2%–4%), and the inner bore retains its integrity, ensuring unobstructed gas pathways. When tension increases (F > 0.20 N), anisotropic deformation occurs, and the cross-section transitions to an elliptical shape due to compressive forces from the yarn. This validates the finite element model’s capability to capture localized stress concentrations. Under critical loading conditions (F ⩾ 0.35 N), structural collapse is evident. Defects include the closure of the inner bore and the formation of folds and pre-cracks on the outer wall. These defects correspond to the regions of high Equivalent Plastic Strain (PE) identified in the simulation. The closure of the inner bore is particularly detrimental, as it directly restricts the gas flow channel, mechanically compromising the functional core of the ECMO device.

Pore structure and gas exchange analysis

The ultimate impact of knitting tension on the functional performance of the ECMO fabric was quantified through porosity analysis, as shown in Figure 7. The relationship between yarn tension and pore structure evolution explains the mechanism of gas exchange efficiency loss.

Figure 7.

PMP membrane porosity test: (a) testing experimental procedure and (b) OP and PC results under different applied forces (F).

The unloaded PMP membrane exhibits a baseline performance with an Open Porosity (OP) of 57.33%, providing a high surface area for gas exchange. Consistent with mechanical and morphological findings, a critical degradation point is observed at 0.2 N. When tension exceeds this limit, the OP drops significantly (ΔOP = 9.22%), while the Percentage of Closed Holes (PC) increases by 21.6%. At high tensions (F > 0.35 N), a pore collapse mechanism dominates: the open porosity falls to less than half of its initial value (47.7%), and closed structures reach 72%. This topological alteration is caused by the integration of localized folds into the pore walls (as seen in SEM), effectively sealing off gas pathways.

The integrated simulation and experimental results consistently demonstrate that knitting-induced compressive contact leads to localized stress concentration at the membrane-yarn interface. Once the applied tension exceeds the critical threshold (0.2 N), the stress state surpasses the yield strength of the PMP membrane, triggering plastic instability, cross-sectional distortion, and pore collapse. This multi-scale damage evolution spanning macroscopic diameter reduction, mesoscopic structural distortion, and microscopic pore closure—collectively explains the deterioration of gas transport performance.

Conclusions

This study elucidates the knitting-induced damage mechanisms of PMP hollow fiber membrane fabrics through a combined numerical and experimental approach. A critical yarn tension threshold of approximately 0.2 N was identified, beyond which irreversible plastic deformation, lumen collapse, and pore closure occur. Below this threshold, deformation remains predominantly elastic and structural integrity is preserved. These findings provide quantitative guidance for tension regulation in warp-knitting processes and contribute to the optimization of low-damage manufacturing strategies for high-performance ECMO membrane fabrics.

Footnotes

ORCID iD

Lifeng Xi

Funding

The authors received no financial support for the research, authorship, and/or publication of this article.

Declaration of conflicting interests

The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

References

Badulak

Antonini

Stead

, et al. Extracorporeal membrane oxygenation for COVID-19: updated 2021 guidelines from the extracorporeal life support organization. ASAIO J 2021; 67(5): 485–495.

Karagiannidis

Brodie

Strassmann

, et al. Extracorporeal membrane oxygenation: evolving epidemiology and mortality. Intensive Care Med 2016; 42(5): 889–896.

Olson

Murphree

Zonies

, et al. Thrombosis and bleeding in extracorporeal membrane oxygenation (ECMO) without anticoagulation: a systematic review. ASAIO J 2021; 67(3): 290–296.

Pappalardo

Schulte

Pieri

, et al. Concomitant implantation of impella® on top of Veno-arterial extracorporeal membrane oxygenation may improve survival of patients with cardiogenic shock. Eur J Heart Fail 2017; 19(3): 404–412.

Makdisi

Wang

IW.

Extra Corporeal Membrane Oxygenation (ECMO) review of a lifesaving technology. J Thorac Dis 2015; 7(7): E166–E176.

Ostadal

Rokyta

Karasek

, et al. Extracorporeal membrane oxygenation in the therapy of cardiogenic shock: results of the ECMO-CS randomized clinical trial. Circulation 2023; 147(6): 454–464.

, et al. Membranes for extracorporeal membrane oxygenator (ECMO): history, preparation, modification and mass transfer. Chin J Chem Eng 2022; 49: 46–75.

Barbaro

Odetola

Kidwell

, et al. Association of hospital-level volume of extracorporeal membrane oxygenation cases and mortality. Analysis of the extracorporeal life support organization registry. Am J Respir Crit Care Med 2015; 191(8): 894–901.

De Somer

. Does contemporary oxygenator design influence haemolysis? Perfusion 2013; 28(4): 280–285.

10.

Hendrix

RHJ

Ganushchak

Weerwind

PW.

Contemporary oxygenator design: shear stress-related oxygen and carbon dioxide transfer. Artif Organs 2018; 42(6): 611–619.

11.

Lachaux

Hwang

Arouche

, et al. A compact integrated microfluidic oxygenator with high gas exchange efficiency and compatibility for long-lasting endothelialization. Lab Chip 2021; 21(24): 4791–4804.

12.

Gehron

Bandorski

Mayer

, et al. The impact of recirculation on extracorporeal gas exchange and patient oxygenation during veno-venous extracorporeal membrane oxygenation-results of an observational clinical trial. J Clin Med 2023; 12(2): 416.

13.

Chotimol

Lansdowne

Machin

, et al. Hypobaric type oxygenators - physics and physiology. Perfusion 2025; 40(2): 273–282.

14.

Scaramuzzo

Gamberini

Tonetti

, et al. Sustained oxygenation improvement after first prone positioning is associated with liberation from mechanical ventilation and mortality in critically ill COVID-19 patients: a cohort study. Ann Intensive Care 2021; 11(1): 63.

15.

Asija

Fried

Siddall

, et al. How I manage differential gas exchange in peripheral venoarterial extracorporeal membrane oxygenation. Crit Care 2023; 27(1): 408–414.

16.

Costa

Halfwerk

Thiel

, et al. Effect of hollow fiber configuration and replacement on the gas exchange performance of artificial membrane lungs. J Membr Sci 2023; 680: 121742.

17.

Ficial

Vasques

Zhang

, et al. Physiological basis of extracorporeal membrane oxygenation and extracorporeal carbon dioxide removal in respiratory failure. Membranes 2021; 11(3): 225.

18.

Dropco

Philipp

Foltan

, et al. Extended cellular deposits on gas exchange capillaries are not an indicator of clot formation: analysis of different membrane oxygenators. ASAIO J 2023; 69(4): e134–e141.

19.

Gimbel

Hsiao

Kim

, et al. A high gas transfer efficiency microfluidic oxygenator for extracorporeal respiratory assist applications in critical care medicine. Artif Organs 2021; 45(8): E247–E264.

20.

Jiang

Jia

, et al. Fluid simulation analysis of ECMO hollow-fiber membrane fabrics from knitting technology. J Ind Text 2024; 54: 1–21.

21.

Wang

, et al. Modification strategies to improve the membrane hemocompatibility in extracorporeal membrane oxygenator (ECMO). Adv Compos Hybrid Mater 2021; 4(4): 847–864.

22.

Wang

Sun

, et al. Improvement hemocompatibility of antithrombotic PMP hollow fiber membranes based on the modification of zwitterionic polymer in ECMO application. J Membr Sci 2024; 704: 122891.

23.

Liu

Bai

Liu

, et al. Exploration of modified Polymers of Intrinsic Microporosity (PIM) for extracorporeal membrane oxygenator (ECMO). J Membr Sci 2025; 717: 123538.

24.

Sun

Zhi

Cheng

, et al. Application of hydrophilic and low-surface-energy anti-fouling membrane for ECMO. J Membr Sci 2024; 701: 122770.

25.

Fukuda

Evolutions of extracorporeal membrane oxygenator (ECMO): perspectives for advanced hollow fiber membrane. J Artif Organs 2024; 27(1): 1–6.

26.

Tang

Zhang

, et al. A novel tunable polypropylene hollow fiber membrane with gradient structure for extracorporeal membrane oxygenation applications. J Membr Sci 2024; 693: 122325.

27.

Zhong

Wang

, et al. Effect of β-cyclodextrin on the hemocompatibility of heparin-modified PMP hollow fibrous membrane for extracorporeal membrane oxygenation (ECMO). Med Nov Technol Devices 2023; 18: 100229.

28.

Alentiev

Evseev

Matson

, et al. Hemocompatibility of promising for ECMO high permeable polyacetylenes. Membr Membr Technol 2024; 6(2): 63–70.

29.

Fukuda

Tanaka

Sadano

, et al. Insights into gradient and anisotropic pore structures of capiox® gas exchange membranes for ECMO: theoretically verifying SARS-CoV-2 permeability. Membranes 2022; 12(3): 314.

30.

Kyosev

Topological modelling of knitted structures. 2019.

31.

Zhuo

Leaf

GAV

Harlock

SC.

The geometry of weft-inserted warp-knitted fabrics part I: models of the structures. J Text 1991; 82: 361–371.

32.

Jia

, et al. Research progress of extracorporeal membrane oxygenation technology in China. J Text Res 2024; 45(3): 234–240.

33.

Jiang

, et al. Low damage preparation of extracorporeal membrane oxygenation warp knit membrane fabrics with adaptive tension. J Text Res 2024; 45(1): 1–9.

34.

Yang

Gao

, et al. Green citrate-based binary diluents for blood oxygenation membranes with high gas permeability. J Membr Sci 2023; 687: 122101.

35.

Sheng

Zhang

Yang

, et al. Preparation and modification of membranes for extracorporeal membrane oxygenator: a review. Chin J Chem Eng 2025; 04: 31–47.

36.

Wang

Liu

Han

, et al. A novel poly (4-methyl-1-pentene)/polypropylene (PMP/PP) thin film composite (TFC) artificial lung membrane for enhanced gas transport and excellent hemo-compatibility. J Membr Sci 2022; 649: 120359.

37.

Guo

Shao

Zhang

, et al. Modified poly (4-methyl-1-pentene) membranes by surface segregation for blood oxygenation. J Membr Sci 2023; 678: 121695.

38.

Dong

Jin

, et al. Hotspots and emerging trend analysis of extracorporeal membrane oxygenation complication research: a qualitative systematic review of literature. Med Nov Technol Devices 2024; 24: 100342.

39.

Markova

Gries

Teplyakov

VV.

Poly(4-methyl-1-pentene) as a semicrystalline polymeric matrix for gas separating membranes. J Membr Sci 2020; 598: 117754.

40.

Jia

Liu

, et al. Research on the PID series control system of a weft-insertion warp knitting machine based on kinematics. J Text Inst 2025; 116: 2761–2771.

41.

Mutschler

Weber

Bueno

MA.

Process analysis and outlook for the development of a new weft yarn inlay system for warp knitting. J Eng Fibers Fabr 2022; 17: 1–13.

42.

Pieklak

Directionally oriented reinforcements of warp-knitted fabrics for composite preforms. Materials 2024; 17(21): 5221.

43.

Hahn

Zierold

Golla

, et al. 3D textiles based on warp knitted fabrics: a review. Materials 2023; 16(10): 3680.

44.

Chen

Low-velocity impact properties of polyurethane composites reinforced with warp-knit spacer fabric. AATCC J Res 2021; 8(2): 38–47.

45.

Liu

Jiang

Dong

A shortened development process method for warp-knitted yarn-dyed shirt fabric. Text Res J 2021; 91(3-4): 443–455.

46.

Liu

Jiang

Dong

Geometric simulation for warp-knitted tubular bandages with the mesh model. Text Res J 2021; 91(21-22): 2612–2623.