Highly curved and flexible SiO₂ fiber sponge via turbulence spinning for efficient and portable plasma/serum separation

Abstract

Blood analysis is essential for managing critically ill patients, yet plasma/serum separation in point-of-care testing (POCT) remains centrifugation-dependent, hindering utility in resource-limited settings. Conventional filtration materials suffer from poor mechanical strength, fragile structures, and suboptimal separation performance. Herein, we fabricate a silica microfiber sponge (Turb-SiO₂) with a 3D disordered network via a turbulence spinning technique. The material exhibits exceptional flexibility, retaining >64% initial stress after 2500 compression cycles, and superior breathability (99.51% air permeability, 0.12 mm H₂O pressure drops) under compression. The Turb-SiO₂ achieves efficient plasma separation with filtrate hemoglobin at 15.7 mg dL⁻¹, matching centrifugation efficacy and outperforming commercial filters. Its crimped fiber structure minimizes hemolysis by eliminating sharp tips that can damage red blood cells. With advantages in weight, volume, cost and operability, the Turb-SiO₂ offers a high-performance solution for portable low-cost POCT blood separation.

A highly curved, flexible SiO₂ fiber sponge produced by turbulence spinning enables rapid, portable plasma separation from whole blood.

1. Introduction

The point-of-care testing (POCT) market is expanding at a compound annual growth rate exceeding 9%. Its core value lies in providing critical diagnostic capabilities for intensive care, disaster medicine, and resource-limited settings, thereby significantly shortening diagnosis time and optimizing clinical decision-making.^1–3 Blood analysis serves as a cornerstone of POCT, as it provides critical physiological and biochemical information that guides therapeutic interventions.^4,5 However, the vast majority of POCT detection technologies are incapable of performing direct analysis on whole blood, rendering the prerequisite plasma/serum separation step indispensable.^6,7

Over the past decade, microscale blood plasma separation strategies have included passive hydrodynamic and filtration-based methods, active acoustophoretic, dielectrophoretic, and magnetophoretic approaches, as well as centrifugal microfluidic platforms.^8,9 Recently, the field has embraced emerging innovations with AI-driven flow control and rapid 3D-printed device fabrication.¹⁰ Despite these technological advances, conventional benchtop centrifuges remain the dominant tool for plasma separation in clinical practice. They are bulky, typically weighing over 20 kg, require a stable power supply, and are ill-suited for true point-of-care deployment in resource-limited or field settings.^11,12 Recognizing this gap, several low cost, lightweight centrifugation alternatives have been developed to address the portability and accessibility challenges.^13,14 Haque et al. benchmarked multiple such alternatives, such as hand powered eggbeaters, 3D printed adaptors, and fidget spinner-based devices. They demonstrated that a $15 3D printed rotor adaptor could process 9 mL clinical blood tubes in a $125 microcentrifuge, achieving >95% plasma yield with hemolysis below 1.7%, well within the clinically accepted threshold of <2%.¹⁵ Despite these notable advances, lightweight centrifuges still require mechanical moving parts, manual operation such as hand-spinning, or external power, and they do not eliminate the need for subsequent plasma transfer to a detection module. Moreover, the cost advantage of a centrifuge is realized only under high-throughput, centralized laboratory conditions, whereas disposable filters become economically favorable in low-volume, resource-limited POCT scenarios.

Membrane filtration technology has emerged as a promising alternative to centrifugation due to its operational simplicity and lack of reliance on complex equipment.^16–21 The core principle of membrane-based blood separation is to retain blood cells while enabling rapid plasma permeation through a porous matrix with precisely controlled pore sizes.^22–25 However, commercially available separation membranes such as glass fiber or polymer membranes exhibit limitations. These materials often present a trade-off between mechanical strength and filtration efficiency. Increasing material density to achieve sufficient mechanical strength leads to higher fluid resistance and pressure drop, requiring greater driving pressure, which in turn elevates the risk of red blood cell rupture and hemolysis.^26–28 For example, Xiao et al. demonstrated that melt-blown nonwoven filtration materials face an inherent conflict between mechanical properties and filtration efficiency when tuning structural parameters. Increasing the basis weight to improve mechanical support results in denser fiber packing, lower porosity, and a significantly higher pressure drop. Their data showed that increasing the basis weight from 24.75 g m⁻² to 32.76 g m⁻² raised the pressure drop to 25.58 ± 12.89 Pa. Maintaining a high filtration efficiency (97.81 ± 0.58%) within a critical basis weight range of 27–29 g m⁻² required even greater driving pressure to overcome the fluid resistance.²⁹ Such high-pressure conditions amplify shear stress on red blood cells (RBCs), increasing the likelihood of cell rupture. Moreover, rigid fiber tips in conventional membranes-particularly in randomly arranged short-fiber structures-physically damage RBCs during filtration. Liu et al. reported that glass fiber membranes induce significant mechanical hemolysis, resulting in filtrate hemoglobin concentrations (Hgb) of 126–268 mg dL⁻¹.³⁰ This severe hemolysis interferes with the accurate quantification of plasma biomarkers, compromising the reliability of downstream biochemical analyses.³¹

Recognizing these limitations, researchers have explored various strategies to improve membrane-based blood separation. Early efforts focused on polymer modification: Venault et al. prepared PEGylated polyvinylidene fluoride membranes via vapor-induced phase separation, achieving a hemolysis ratio below 1% and extended plasma clotting time.³² Subsequently, Salehi Shahrabi et al. fabricated blended fibrous mats by electrospinning poly(ethylene terephthalate) with polyvinylpyrrolidone; the addition of 5 wt% PVP yielded uniform fiber morphology and platelet permeation above 90%, while leukocyte removal increased with the number of membrane layers.³³ More recently, Wang et al. developed a zwitterionic polyurethane-modified cellulose acetate membrane that reduced fibrinogen adsorption by 60% and enabled rapid plasma extraction, yielding 0.5–0.7 mL of plasma from 10 mL of whole blood within 10 minutes under vacuum-driven filtration.⁶ In another effort, Dizon et al. integrated vapor-induced phase separation with electrospinning to fabricate asymmetric cellulose acetate composite membranes, achieving complete retention of platelets, red blood cells, and white blood cells for five-fold diluted whole blood under gravity-driven filtration.³⁴

Despite these advances, several critical challenges persist. First, most polymer-based membranes still rely on blood dilution to achieve acceptable filtration performance, as undiluted whole blood—characterized by high viscosity and cellular density—often leads to rapid clogging and flow cessation.^26–30 Second, achieving high cell retention frequently comes at the cost of increased hemolysis or reduced throughput, a trade-off inherent to the limited mechanical robustness and pore structure stability of polymeric materials.^32,33 Third, the fabrication processes for advanced membranes—such as vapor-induced phase separation or electrospinning with precise parameter control—remain relatively complex and may hinder scalable production and cost-effective deployment in resource-limited settings. Clinical standards for plasma quality demand not only high cell retention, typically less than 0.1% residual red blood cells, but also stringent control of hemolysis: free hemoglobin should remain at a low level to avoid interference with colorimetric or immunoassays. The root of these challenges lies in the difficulty of traditional materials to simultaneously optimize multiple key parameters: high porosity,³⁵ excellent mechanical elasticity,³⁶ precise pore size control,³⁷ and inherent biocompatibility.³⁸ Therefore, there is an urgent need for a novel filtration material that can concurrently achieve rapid flow, sufficient mechanical robustness, high plasma yield, and ultra-low hemolysis—a requirement that demands innovation in both fiber architecture and fabrication methodology.

Recent advances in spinning technologies have enabled precise control over micro/nanofiber architectures.^39,40 Traditional blow spinning SiO₂ fiber membrane achieves sub-micron fiber diameters but suffers from poor elasticity due to their 2D laminated structure, retaining only 34% of initial stress after 2500 compression cycles-limiting their reliability in plasma/serum separation. In contrast, turbulence spinning techniques leverage fluid instability to induce fiber crimping and entanglement, forming mechanically robust 3D disordered networks; For example, Li et al. developed a Kármán vortex street solution blow spinning method, where high-speed airflow generates fluid oscillations that drive 3D jet movement, resulting in highly crimped, interwoven fibers with sponge-like porous structures.⁴¹ This architecture combines low density, high resilience, and continuous large pores-properties ideally suited for filtering shear-sensitive biofluids such as blood. However, the application of turbulence spinning to fabricate rigid inorganic oxide fibers and systematic evaluation of their comprehensive performance for whole blood separation, including mechanical durability, air permeability, and hemolysis control, remain underexplored.

Herein, we propose and validate an integrated turbulence spinning strategy for fabricating a silica microfiber sponge (Turb-SiO₂) with a 3D highly crimped network structure that exhibits greater fiber curl than conventional blow spinning counterparts. We systematically investigated the effects of turbulence spinning parameters on the micromorphological properties of as-prepared Turb-SiO₂. The Turb-SiO₂ demonstrates exceptional comprehensive performance, including a low density of 0.14–0.18 g cm⁻³, high cyclic stability with over 64% of initial stress retained after 2500 compression cycles at 60% strain, high air permeability of 99.51%, and a low-pressure drop of 0.12 mm H₂O. Moreover, blood filtration experiments confirm that the filter reduces the filtrate Hgb to 15.7 mg dL⁻¹ within 4 minutes, demonstrating not only a high-performance solution to the “last mile” challenge of blood separation in POCT, but also the potential of turbulence spinning for fabricating functional fibers in biomedical engineering.

2. Experimental procedures

2.1. Materials

Polyvinyl alcohol (PVA, M_w 205 000), poly(ethylene oxide) (PEO, M_w 400 000), and phosphoric acid were purchased from Aladdin. Tetraethyl orthosilicate (TEOS) was purchased from Macklin.

2.2. Characterization and plasma/serum separation tests

The overall dimensions of the SiO₂ fiber sponges were measured using a digital vernier caliper with an accuracy of 0.01 mm. The microstructure and morphology of the samples were observed using a scanning electron microscope (SEM, Carl Zeiss AG Sigma 500). Fiber diameter distributions and internal pore sizes within the sponges were statistically analyzed from the SEM images using ImageJ software. Phase analysis was performed using a Rigaku Ultima IV X-ray diffractometer with Cu Kα radiation (λ = 1.5406 Å). The chemical states and elemental composition were characterized by X-ray photoelectron spectroscopy (XPS, Thermo Scientific Escalab Xi⁺). Thermal stability and compositional changes of the precursor fibers were investigated via thermogravimetric analysis (TGA) obtained by NETZSCH STA 449 F3/METTLER 1100LF instrument over a temperature range of 25 °C to 1300 °C. The air permeability was measured using a CERTITEST 8130A automated filter tester (TSI Inc.). Static contact angles were measured via the pendant drop technique. A 2 µL droplet was extruded from a needle moving downward at a constant rate. Upon contact of the droplet with the sample surface, the needle was quickly retracted, and the contact angle was immediately captured from the image at the instant of retraction. The angle values were obtained by fitting the droplet shape with the Young–Laplace equation. Fourier Transform Infrared (FT-IR) spectroscopy was conducted using a Nicolet is 20 spectrometers with a scanning range of 4000 to 400 cm⁻¹ to analyze the chemical functional groups and bonding characteristics of the materials. Compression tests were performed using a computer-controlled universal testing machine (model C42.503). Tests included single compression to various strain levels and cyclic compression, both conducted at a constant rate of 60 mm min⁻¹. For the cyclic tests, samples were subjected to 3000 compression cycles at 60% strain. Blood filtration under negative pressure was performed using a FluidicLab Precision Pressure Pump/Controller for Microfluidics (Model PC1). The device was set to deliver a constant negative pressure of −140 mbar as used in the plasma separation experiments.

The blood specimens were provided by Jinan Central Hospital Affiliated to Shandong First Medical University. All sample collections were approved by the Ethics Committee of the Jinan Central Hospital Affiliated to Shandong First Medical University (JIKELUNSHEN: 20231226022), and the study was conducted in accordance with relevant guidelines. As the clinical gold standard control, whole blood was centrifuged at 3000×g for 15 minutes to obtain plasma/serum. Hgb in the filtered serum was measured using a hemoglobin colorimetric assay kit (Beyotime Biotechnology, Shanghai, China, Catalog No. P0381S), following the manufacturer's instructions. Blood samples were collected from eight healthy volunteers (four males and four females) under ethical approval. Prior to filtration, each sample was analyzed using a hematology analyzer (Sysmex XP-300) to determine hematocrit (Hct) and red blood cell (RBC) count. For blood filtration testing, each blood sample with a volume of 1 mL was filtered through the Turb-SiO₂ membrane under gravity. The same blood samples were centrifuged at 3000 rpm for 10 minutes using an Eppendorf 5417R centrifuge as the control. Free hemoglobin concentration in the filtrate and centrifugate was determined by the Cripps method based on spectrophotometric measurements at 560, 576 and 592 nm. Residual red blood cell count in the plasma was measured using the same hematology analyzer with a detection limit of approximately 5 × 10⁴ L⁻¹. Cell removal efficiency was calculated as the difference between the initial and residual red blood cell counts divided by the initial count, multiplied by 100%. Hematoxylin and eosin (HE) staining utilized a tissue processor (DIAPATH Donatello), embedding machine (Wuhan Junjie JB-P5), microtome (Shanghai Leica RM2016), freezing stage (Wuhan Junjie JB-L5), tissue spreader (Jinhua Kedi KD-P), and microscope with imaging (Nikon E100 + DS-U3). Digital scanning used a 3DHISTECH Pannoramic scanner and CaseViewer 2.4.

3. Results and discussion

3.1. Synthesis and characterization

The laminar and turbulent air flows were generated using a straight uniform tube and an orifice with a sharp-edged contraction, respectively. These two configurations produced distinctly different flow states, which in turn led to marked differences in the morphology and curvature of the resultant fibers. Fibers produced under laminar flow were straight and orderly, whereas those formed under turbulent flow were bent and disordered (Fig. 1a). The airflow spinning system consisted of an air compressor, a drying tube, a syringe pump, a needle, a flow outlet assembly, and a collection cage (Fig. 1b). Interchangeable outlet assemblies, consisting of a straight uniform metal tube (Fig. 1c) and a rapidly contracting nozzle (Fig. 1d), enabled active control over the laminar and turbulent flow regimes. The precursor nanofibers, ejected by the turbulent airflow, were captured by a porous cage and assembled into a three-dimensional sponge-like preform. Its macroscopic appearance and microstructure are shown in Fig. 2a. Measurements indicated an average diameter of 8.5 cm for this precursor sponge. Subsequently, the preform was densified, then calcined in air to remove the PVA and PEO organic components and to induce crystallization, yielding the final Turb-SiO₂ microfiber sponge. Following calcination, the organic PVA and PEO were completely removed, resulting in a pure SiO₂ microfiber sponge (Fig. 2b). A noticeable shrinkage was observed, reducing the average sponge diameter to 7.8 cm. The resulting SiO₂ sponge exhibited a very low density (Fig. 2c), ranging from 0.14 to 0.18 g cm⁻³. This is significantly lower than the density of 0.20 to 0.28 g cm⁻³ for comparable samples produced by blow spinning. This difference originates from the distinct fiber formation mechanisms. The turbulence spinning process employs a high-speed turbulent field to stretch and deposit fibers. This results in looser fiber packing, creating an open, disordered three-dimensional network with larger pores and a lower solid fraction per unit volume. In contrast, fibers in blow spinning follow more linear trajectories, leading to denser packing upon deposition, which yields a higher packing density and lower porosity. A comparison of the macro- and micro-structures of precursor fibers from blow spinning (Fig. 2d) and turbulence spinning (Fig. 2e) confirms that the turbulent process imparts a significantly more crimped morphology to the fibers. To investigate the effect of airflow velocity on fiber diameter, SiO₂ sponges were fabricated at turbulent airflow velocities of 6, 8, and 10 m s⁻¹. Statistical analysis of fiber diameters from the corresponding SEM images (Fig. 2f) yielded the distributions shown in Fig. 2g. The average fiber diameters were 2.0 ± 1.0 µm, 1.2 ± 0.6 µm, and 1.0 ± 0.5 µm for velocities of 6, 8, and 10 m s⁻¹, respectively, demonstrating that higher airflow velocities produce finer fibers. Finally, the elemental spectrum of the calcined fibers (Fig. 2h) confirms the complete conversion of the precursor to SiO₂.

Fig. 1. Schematic illustration and experimental setup of the blow spinning and the turbulence spinning systems. (a) Airflow regime visualization and corresponding fiber morphology; (b) integrated experimental setup of the spinning system; (c) image of the blow spinning gas for blow spinning; (d) image of turbulence gas for turbulence spinning.

Fig. 2. Structure, morphology and properties of Turb-SiO₂. (a) Macroscopic morphology and microstructure of precursor fiber sponge; (b) macroscopic morphology of as-prepared Turb-SiO₂, and (c) a demonstration of its ultra-low density; (d) microstructure of the blow spinning SiO₂ fibers, and (e) Turb-SiO₂; (f) micro-morphology of Turb-SiO₂ prepared at different turbulent airflow velocities; (g) statistical diameter distribution of SiO₂ fibers prepared at different turbulent airflow velocities; (h) elemental mapping (C, O, Si) of Turb-SiO₂.

3.2. Compression test of SiO₂ fiber sponges at room temperature

The Turb-SiO₂ microfiber sponge exhibits excellent flexibility and elasticity. As shown in Fig. 3a, the sponge can withstand repeated compression and recover its original shape after cyclic loading. To investigate its compressive properties, uniaxial compression tests were conducted at room temperature on both the Turb-SiO₂ sponge and a traditional blow spinning SiO₂ fiber mat, up to 60% strain. The results (Fig. 3b) show that after 500 cycles of structural conditioning, the maximum stress for the Turb-SiO₂ sponge at 60% strain was 3.72 kPa, compared to 11.87 kPa for the traditional sample. The stress evolution during 2500 compression cycles under 0 to 60% strain, following the initial 500 conditioning cycles, is presented in Fig. 3c and d. After 2500 cycles, the compressive stress of the Turb-SiO₂ sponge at 60% strain remained above 64% of its initial value of approximately 3.72 N, demonstrating excellent cyclic stability. In contrast, the blow spinning mat retained only 34% of its initial stress of approximately 11.87 N after the same number of cycles, indicating significantly inferior resilience under frequent compression. This performance difference is attributed to distinct fiber architecture. The disordered three-dimensional network formed by turbulence spinning distributes external forces more uniformly throughout the material, preventing stress concentration and creating a more robust structure. Conversely, the more aligned and ordered fibers in the blow spinning mat are prone to premature localized failure. As illustrated in Fig. 3e, under moderate loads ranging from 1 to 7 N, only localized fiber breakage occurs, which does not significantly affect the overall compression and recovery. However, under higher loads of 7 to 14 N, widespread fiber fracture occurs, transforming the structure into shorter, straighter fragments and leading to a marked decline in resilience. The compression test results confirm the superior elastic performance of the Turb-SiO₂ microfiber sponge.

Fig. 3. Compressive performance and structural evolution of SiO₂ fiber-based materials. (a) Demonstration of the cyclic compression-recovery process of the Turb-SiO₂; (b) uniaxial compression stress–strain curves of the Turb-SiO₂ and blow spinning SiO₂ fiber mat at 60% strain; (c) stress evolution of the Turb-SiO₂ during 2500 cyclic compression (0–60% strain) cycles; (d) stress evolution of the blow spinning SiO₂ sponge during 2500 cyclic compression; (e) microstructural evolution of the Turb-SiO₂ under different loads.

3.3. Breathability, material modification and biocompatibility

The high air permeability, low air resistance, and hydrophilicity of Turb-SiO₂ make it highly suitable for blood filtration under a negative pressure of −140 mbar. Air permeability tests (Fig. 4a) show that at an airflow rate of 32.1 L min⁻¹, the material exhibits a pressure drop of only 0.12 mm H₂O and a permeability of 99.51%. Even under a compressive stress of 6 N for densification, the permeability of Turb-SiO₂ remains at 95.02%. Furthermore, the pressure drops increase linearly with compression force (R² = 0.97), confirming the excellent compressive stability of the three-dimensional microfiber network (Fig. 4b). After surface modification via plasma treatment, the water contact angles for both the Turb-SiO₂ and the blow spinning SiO₂ fiber membrane were significantly reduced (Fig. 4c), indicating improved surface wettability and enhanced hydrophilicity. These characteristics ensure stable filtration performance under negative pressure (Fig. 4d). During blood filtration, even under fluid shear stress, the densified fiber network and the red blood cells retain their structural integrity. Blood separation tests using the Turb-SiO₂ (Fig. 4e) were conducted, and the filter samples were examined post-filtration via HE staining to observe red blood cell integrity (Fig. 4f). The images show that the long, continuous fibers in the Turb-SiO₂ sponge help preserve red blood cell integrity, as the fibers lack exposed sharp tips that could puncture cells and cause hemolysis (Fig. 4g). In contrast, short, straight fibers tend to expose numerous sharp tips, leading to cell rupture (Fig. 4h). These results demonstrate the good biocompatibility of the Turb-SiO₂ and its ability to significantly reduce the risk of large-scale hemolysis during filtration.

Fig. 4. Performance characterization and blood filtration application of Turb-SiO₂ microfiber sponges. (a) Curves of permeability and flow rate of the Turb-SiO₂ under different densification pressures; (b) pressure drop variation of the Turb-SiO₂ under different densification pressures; (c) comparison of water contact angles between Turb-SiO₂ and conventional blow spinning SiO₂ fiber mat before and after plasma treatment; (d) schematic diagram of the experimental setup for blood filtration under negative pressure (−140 mbar); (e) experimental process of blood separation using Turb-SiO₂; (f) macroscopic morphology of the Turb-SiO₂ after blood filtration; (g) micro-morphology of red blood cells after filtration by Turb-SiO₂; (h) micro-morphology of red blood cells after filtration by traditional short straight fibers.

3.4. Blood filtration performance

The pore-size-adapted Turb-SiO₂ demonstrates excellent blood filtration performance and high cost-effectiveness. The pore sizes of different filter samples were statistically analyzed from SEM images, where the inter-fiber voids were fitted as ellipses and the major axis was taken as the pore diameter. Analysis of four distinct filters revealed that the fabrication method, densification pressure, and spinning airflow velocity all significantly influence the pore structure. Fig. 5a shows four types of SiO₂ fiber filters fabricated via different manufacturing processes, including blow spinning at an air velocity of 8 m s⁻¹, and turbulence spinning at the air velocities of 6, 8, and 10 m s⁻¹. Generally, a smaller average pore size correlated with a clearer filtered serum sample. Whole blood filtration tests were conducted using filters produced under different conditions, and the Hgb in the filtered serum was quantified (Fig. 5b). For filters made via turbulence spinning, the results varied with airflow velocity: filters produced at 6 m s⁻¹ showed no effective filtration. Those produced at 10 m s⁻¹ yielded serum with an average Hgb of 87.63 mg dL⁻¹, with the best result of 76.71 mg dL⁻¹ achieved at a densification pressure of 3.5 N, indicating modest filtration efficacy. In contrast, filters produced at 8 m s⁻¹ performed exceptionally well, yielding serum with an average Hgb of 41.25 mg dL⁻¹ and a minimum value of 15.7 mg dL⁻¹ at 3.5 N densification pressure. A direct comparison confirms that filters fabricated by turbulence spinning consistently outperform those made by blow spinning, producing serum with lower residual hemoglobin (Fig. 5c). The performance was further benchmarked against the clinical standard (centrifugation) and a commercial SiO₂ fiber membrane filter (Fig. 5d and e). The Hgb in serum filtered by the commercial membrane was more than double that found in samples processed by either centrifugation or Turb-SiO₂ filter. Notably, the hemoglobin level of 15.7 mg dL⁻¹ achieved by our optimized filter approached that of centrifugation. In clinical diagnostics, a low serum Hgb level is considered acceptable, as it does not significantly interfere with subsequent analyses. The hematological parameters are summarized in Table 1. The Hct ranged from 33.4% to 48.6% and RBC counts from 4.13 to 5.35 × 10¹² L⁻¹, covering normal physiological ranges. After filtration, the maximum free hemoglobin concentration in the filtrate was 42.0 mg dL⁻¹ for sample 3, which falls below the interference thresholds of all commercial IVD kits listed in Table S1. These results confirm that the Turb-SiO₂ membrane produces plasma of sufficient quality for clinical sample preparation and downstream diagnostic applications. A comprehensive comparison of key parameters, including mass, volume, cost, processing time, and filtration efficacy, is summarized in Fig. 5f. Direct purchase-price comparison between a durable centrifuge and our single-use filter would be inappropriate; we instead adopt a cost-per-test framework following Haque et al.¹⁵ As shown in Fig. 5f, for a low-throughput scenario of 100 tests per year, the amortized cost of a standard benchtop centrifuge exceeds $10 per test, while a low-cost modified centrifuge still costs approximately $1.5 per test. In contrast, our Turb-SiO₂ filter requires no capital investment, electricity, or maintenance, with a per-test cost below $1. The cost of our Turb-SiO₂ membrane is comparable to that of commercial SiO₂ fiber membrane filters, yet our filter offers superior separation quality with significantly lower hemolysis levels. Centrifugation remains cost-effective in high-throughput centralized laboratories, but our filter offers superior portability, speed, and separation quality for decentralized or field-based applications. The detailed cost assumptions and calculation parameters are provided in Table S2 in the SI, with the key per-test cost comparison summarized in Fig. 5f.

Fig. 5. Pore structure, blood filtration performance and parameter comparison of Turb-SiO₂ and other blood separation methodologies. (a) Microscopic morphology and pore size distribution of filters under different preparation conditions; (b) Hgb in filtered serum of Turb-SiO₂ filters under different densification pressures and turbulent airflow velocities; (c) comparison of Hgb in filtered serum between turbulence spinning and blow spinning filters under different densification pressures; (d) comparison of Hgb in filtered serum among Turb-SiO₂ filter, centrifugation (clinical standard) and commercial SiO₂ fiber membrane filter; (e) comparison of experimental setups of centrifugation, commercial SiO₂ fiber membrane filter and Turb-SiO₂ filter; (f) comprehensive comparison of key parameters including mass, volume, per-test cost based on a low-throughput scenario of 100 tests, processing time, and filtration efficacy in terms of filtrate hemoglobin among centrifugation as the clinical standard, a commercial SiO₂ fiber membrane filter, and the Turb-SiO₂ filter.

Table 1. Hematological parameters of eight donor blood samples before and after filtration using the Turb-SiO₂ membrane, with centrifugation control^a.

Sample ID	Sex	Hct (%)	RBC count (×1012 L−1)	Hgb (mg dL−1)		Residual RBC count (×104 L−1)
		Pre-filtration	Pre-filtration	Filtrate	Centrifugate	Filtrate	Centrifugate
1	F	42.1	4.7	22.9	0.1	90	140
2	M	48.6	5.35	14.2	0.3	20	210
3	F	41.4	4.43	42	0.3	240	200
4	F	33.4	5.34	25.1	2.1	1060	340
5	M	42.9	4.51	13	0.1	1900	130
6	M	45.3	4.94	19.6	0	1240	40
7	F	37.8	4.16	16.6	8.8	1420	30
8	F	37.8	4.13	27	1.3	705	220

^aReference normal ranges: hematocrit: male 40–50%, female 36–46%, RBC count: male 4.0–5.9 × 10¹² L⁻¹, female 3.8–5.2 × 10¹² L⁻¹.

4. Conclusion

In summary, to tackle the technical bottlenecks plaguing POCT, particularly the inadequate mechanical strength and inherent fiber brittleness of conventional filtration media, this study adopted an innovative turbulence spinning technique to fabricate ultra-light and flexible Turb-SiO₂, featuring a highly curved 3D network architecture. This distinctive disordered 3D structure confers the material with exceptional elasticity and resilience, accompanied by the favorable intrinsic properties of low density and high porosity. Furthermore, it exhibits critical physical characteristics that are well-suited for blood filtration applications, including prominent air permeability under a low pressure drop. In functional validation experiments, Turb-SiO₂ demonstrated remarkable capability in preserving the integrity of red blood cells. The Turb-SiO₂ filter is lighter, more compact, less expensive, and faster to use than a centrifuge, while outperforming the commercial filter in separation quality. The Hgb level in the filtered serum was as low as 15.7 mg dL⁻¹, which not only indicates a significantly reduced hemolysis risk in comparison to traditional filtration materials but also achieves a separation efficacy that approximates the gold standard of centrifugation. Collectively, these superior performance advantages validate that the Turb-SiO₂ effectively overcomes the issue of fiber brittleness inherent in conventional media and realizes the integration of structural design and functional performance. Consequently, this work lays a robust material foundation for the development of high-efficiency, low-cost, and portable devices dedicated to blood separation and filtration, thereby holding substantial potential for advancing the practical application of POCT technologies in clinical and resource-limited settings.

Author contributions

H. W. and S. L. conceived the idea and designed the experiments. J. F. Z., H. W. and S. L. supervised the project. S. D. Z., Z. K. C., and H. W. prepared materials/production. S. D. Z., Y. G., Z. J. F., S. T. H., Y. Y. Z., Y. H. P. and S. L. characterized micromorphology. H. H. C. performed mechanical tests/analysis. M. Z., Y. W., Q. L. X., and S. L. conducted thermal tests/analysis. S. D. Z., H. H. C., M. Z., Y. W., Q. L. X., R. X, and J. F. Z. executed separation tests/analysis. S. D. Z. handled the HE scan. All authors discussed the results of this work and contributed to the writing and revision of the manuscript.

Conflicts of interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Supplementary information (SI) is available. See DOI: https://doi.org/10.1039/d6na00366d.