Applications for a 3D Printed, Handheld, Low-Cost Centrifuge in Separating Pathogenic Bacteria from Biofluids

ABSTRACT

Centrifugation is an important step in the preparation of biofluids for medical analysis. Traditional centrifuges require electricity and are expensive and difficult to transport, presenting barriers to accurate disease diagnosis in resource-poor communities. While prior research has shown the potential of low-cost centrifuge alternatives, 3D printing is an underexplored approach that could enable large-scale, low-cost production. Further, existing low-cost centrifuges have not tested the separation of bacteria from saliva or the separation of analytes at various viscosities, which are important capabilities for real-world applications. Here, my goal was to design a 3D-printed, low-cost centrifuge and test its effectiveness in compacting pathogenic bacteria in various viscous environments. Using fluorescent E. coli-spiked saliva and micro-fluorescent beads in water and glycerol, the fluorescence intensity of samples pre- and post-centrifugation was measured. Results showed a significant increase in the fluorescence intensity of the centrifuged saliva samples over the non-centrifuged samples (p<0.001). Furthermore, testing done with micro-fluorescence beads in different concentrations of glycerol, mimicking different viscosities, showed a significant separation up to 30% glycerol (p<0.001) but no difference between the separation of 40% and 50% glycerol (p=0.337). Finally, my device showed separation comparable to that of a commercial centrifuge in the range of 1,000-2,000 RPM. These results indicate the potential applications of this device for separating pathogenic bacteria from biofluids, such as saliva of varying viscosities, and the feasibility of 3D printing as an effective means of production. In the future, this device has the potential to provide resource-poor communities with durable, low-cost, and efficient centrifugation.

INTRODUCTION.

Diagnostic tests are an important tool in providing essential information about a patient’s condition and are beneficial towards establishing proper care and treatment for those diagnosed with specific results [1]. Many diagnostic tests examine aspects of a biofluid sample for specific analytes or compounds that can be measured and used in the diagnostic process [2-4]. Many samples used require preparation prior to examination to obtain clearer results. Furthermore, many samples, such as blood, saliva, or urine, require compaction or separation for sample analysis [5]. Compaction increases the concentration of low-abundant analytes, such as bacteria below sub-infectious levels, thereby providing more accurate diagnostic results.

In many cases, centrifugation plays an important role in this process as it concentrates and separates biofluid samples prior to analysis for diagnosis [6]. Traditionally, however, access to centrifugation requires access to expensive, bulky, difficult to transport, and electricity-dependent commercial devices [7-8]. These issues combined pose a significant limitation to areas that lack sufficient resources and infrastructure and suffer from high rates of disease transmission [9]. This issue prompts the need for diagnostic technologies that can prepare samples for diagnostics while also being portable, low-cost, and easy to use [6].

Beginning with the Paperfuge [5] and its paper-based, low-resource, handheld centrifuge design, several models have been developed to address this issue, providing efficient separation of biofluids, primarily blood plasma, for medical applications [8]. Additionally, designs have been adapted for other applications such as drug preparation [10] and the separation of gold nanoparticles [11]. Many of these designs utilize cheap, easily available materials such as paper and cardboard to construct lightweight designs that rotate because of a coiling string being pulled back and forth [5]. Other low-resource centrifugation techniques have also been developed using designs based on salad spinners [12], eggbeaters [13], and fidget spinners [14-15] which have shown promise with the separation of biofluids. Of these models, the string-based design remains the most prominent due to the fastest rotational speeds being demonstrated with this design.

These designs do leave room for variability, however, due to the hand-cut paper and cardboard structures, which can differ between devices and subsequently yield different results upon testing. Furthermore, these designs lack long-term durability due to their inherently weaker components, such as paper and cardboard. These limitations led to the possibility of using 3D printing to create more standardized, durable handheld centrifuge devices [16]. These designs have the added benefit of being easily mass producible while remaining cost-efficient. So far, research has shown the potential of 3D printing as a means of production for low-resource centrifuges by designing and testing 3D-printed devices for nucleotide extraction in DNA sequencing [16].

Furthermore, the separation of the cellular components of saliva has not been tested with a low-cost, electricity-free device. Saliva is an easily obtained biofluid that contains numerous analytes useful for disease diagnosis without requiring invasive techniques like drawing blood for serum analysis [17]. Furthermore, saliva has recently been shown to be comparable to serum in displaying an individual’s health [18]. Analytes can include many different pathogenic bacteria such as S. mutans, H. influenzae, N. gonorrhoeae, and T. pallidum [19]. Due to the large amount of bacterial diversity that exists in saliva [20], and the relatively low concentration of specific pathogenic bacteria, centrifugation is an important step in preparing samples for analysis.

In resource-limited communities, the benefits of saliva testing paired with the low cost and simplicity of handheld centrifugation present a promising opportunity to enhance diagnostic capabilities and accuracy through efficient sample preparation prior to diagnostics.

Prior research on low-cost centrifugation has not yet investigated the potential of these devices for the compaction of pathogenic bacteria in biofluids for disease detection and diagnosis, nor has it examined the feasibility of using saliva as a testing sample for analyte concentration. Jointly, it is possible that these designs can be a useful tool for assisting in the detection of pathogenic bacteria often found in saliva.

Additionally, different biofluids used for diagnostic testing exhibit varying viscosities, and previous research has shown that saliva and blood individually exhibit a wide range of viscosities [21-22]. Furthermore, it has been shown that nanoparticle separation in biofluids differs by viscosity [23]. Thus, viscosity is an important property to consider when looking at biofluid separation. Previous research has yet to examine whether a handheld centrifuge can separate samples across varying viscosities. Thus, an additional aim of this study is to examine sample separation at differing viscosity levels to demonstrate a broader application of this technology.

This study involves the design and implementation of a handheld, low-cost, 3D-printed centrifuge. We demonstrate the application of this design for separating bacteria from saliva samples as proof-of-concept for pathogen identification and diagnosis in saliva. Additionally, we demonstrate the effectiveness of this device at separating micro-fluorescent beads across different viscosities, demonstrating its separation capabilities for a variety of biofluids and saliva samples with different viscosities.

MATERIALS AND METHODS.

The centrifuge designed here is constructed almost entirely of 3D printed tough polylactic acid (PLA) filament with two 3D-printed holders, a central disc that holds the PCR tubes, and nylon strings.  Design work was done using Autodesk Fusion, and the final product was printed using an Ultimaker S5 printer (Figure 1).

Prior to centrifugation, two 0.2 mL PCR tubes are inserted into the cylindrical holes in the centrifuge, ensuring even weight distribution across both sides (Figure 2A). Holding the device with the string relaxed, the user swings the disc forward in a slow circular motion to coil the string. Once the string has been sufficiently coiled, the hand holds are firmly pulled outwards, causing the string to uncoil and the device to rotate. After the string has been completely uncoiled, the hand holds are relaxed, allowing for the device to recoil (B). This motion is then repeated to maintain the device’s rotation [10]. After two minutes of rotation, the device is stopped, and the test tubes are carefully removed, taking care to ensure that no mixing occurs (C). Then, to measure the concentration of the pellet, the noncompacted supernatant is removed (D) and the 20 µL pellet is left for analysis (E). This pellet is then transferred into a Corning 384 low-volume microplate well for analysis (F). Finally, fluorescence spectroscopy measurement is performed to measure the fluorescence intensity of the sample on a Tecan Microplate reader (G).

A baseline test was conducted to evaluate the effectiveness of the handheld device compared to commercially available electric centrifuges. Here, a mixture of micro-fluorescent beads that mimic the size and separation of bacteria with visible concentration was used. First, a stock solution of Fisher-Scientific yellow-green FluorSphere 2-micron beads was diluted using a 1:5 ratio of beads to deionized (DI) water. Additionally, a stock solution of Fisher Scientific red 20 nm beads was used, consisting solely of red beads. The larger yellow-green beads were the variable of interest in this study due to their similarity in size to most pathogenic bacteria, while the red beads were smaller and more representative of small particulate matter found in biofluids such as DNA aggregates or proteins.

To compare the low-cost 3D-printed centrifuge with a commercially available centrifuge, a Fisher Scientific Gusto Mini Centrifuge was used. A 20 µL solution of both the yellow-green bead and 20 µL of the red bead was added to a 0.2 mL PCR tube, along with 160 µL of DI water. These PCR tubes were then spun for 2 minutes using either the handheld centrifuge or the commercial centrifuge set at 1,000, 2,000, or 3,000 RPM (63, 251, and 564 RCF). This process was repeated three times for each centrifuge and setting.

To compare the level of compaction achieved by the samples, 180 µL of the non-compacted supernatant was removed after centrifugation. This left 20 µL of the compacted yellow, fluorescent bead pellet in the PCR tube. This remaining volume was then transferred into a well plate for fluorescence spectroscopy measurements.

To test the effectiveness of the handheld centrifuge in compacting samples of the different viscosities, the same stock solutions of yellow-green and red micro-fluorescent beads were utilized. To test the separation of different viscosities, 20 µL of the red bead solution and 20 µL of the yellow-green bead solution were combined and added to a PCR tube along with 100 µL of the glycerol solution diluted in DI water. The glycerol dilutions ranged from 0% to 50% glycerol with each solution increasing by 10%. The same methodology described above was utilized to compact the samples. This process was repeated three separate times for each dilution.

To test bacteria separation from saliva, Green Fluorescent Protein (GFP) expressing E. coli was used because E. coli is an oral pathogen of interest to the public and because of the GFP portion’s capability for detection using the fluorescence spectrometer. GFP-E. coli was added to saliva at a final concentration of 1.1 x 10⁷ CFU/mL to mimic physiologically relevant concentrations. A control group consisting of non-spiked saliva and water was also created. To ensure an even distribution, both solutions were vortexed for 15 seconds.

Prior to centrifugation, 20 µL of the spiked saliva was taken out of the sample. This sample was then compared to the centrifuged sample to determine the level of compaction gained from testing. To obtain the centrifuged sample, the remaining 200 µL of the saliva were added to a PCR tube and spun for two minutes using the handheld centrifuge. Following centrifugation, 170 µL of supernatant was removed. The remaining pellet was then vortexed prior to collection to ensure complete distribution of the pellet. 20 µL was then removed for fluorescence spectroscopy measurements. The discrepancy in volumes was due to variable compaction that occurred among the saliva samples during centrifugation and the need to keep the volume of analyzed pellets the same across all trials.  The same process was repeated for a non-spiked control group to ensure the fluorescence was a result of the bacteria and not natural fluorescence in the saliva. After fluorescence spectroscopy was performed, the fluorescence intensities of the non-centrifuged and centrifuged samples were compared to determine the level of compaction achieved by the handheld centrifuge.

RESULTS.

A comparison of fluorescence intensity was performed between the handheld device and the Fisher-Scientific commercial centrifuge fitted with a PCR tube adapter (Figure 3). All trials were performed for 2 minutes, with the pellet being analyzed for fluorescence intensity. Averaging the data showed 34.86% separation at 1,000 RPM (63 RCF), 64.13% at 2,000 RPM (251 RCF), and 75.84% at 3,000 RPM (564 RCF) compared to the intensity of the stock solution. The handheld device, when averaging three separate trials, measured 41.25% separation. An unpaired t-test between the fluorescence intensity of the handheld centrifuge and that of the commercial centrifuge at 1,000 RPM showed no significant difference (p=.066).

A similar comparison of fluorescent intensity was performed for the handheld device when separating samples of varying viscosities (Figure 4). When looking at the fluorescence intensity of samples with increasing levels of glycerol and viscosity, there was a decrease in the percentage of beads that were separated as viscosity increased.  A one-way ANOVA on percent separation showed significant differences across all groups (p<0.001), but an unpaired t-test between 40% and 50% glycerol solutions specifically showed no significant difference (p=0.337).

Fluorescence intensity was measured for the GFP-E. coli spiked and non-spiked saliva before and after centrifugation using the handheld device for two minutes (Figure 5; Left). Baseline testing showed a significant increase in fluorescence intensity of the centrifuged saliva samples with GFP-E. coli when compared to the non-centrifuged saliva samples with GFP-E. coli after running an unpaired t-test (p<.001).

Five trials, performed on different days, were conducted to measure fluorescence intensity before and after centrifugation of saliva samples spiked with GFP-E. coli at a concentration of 10⁷ CFU/mL (Figure 5; Right). These trials showed a significant increase in the fluorescence intensity of the centrifuged saliva when compared to the non-centrifuged saliva after running an unpaired t-test on the peak fluorescence intensity of all trials (p=0.03). The mean fluorescence intensity of the non-centrifuged sample was 6,096 a.u., and the mean fluorescence intensity of the centrifuged sample was 10,529 a.u.

Averaging the concentration of the fluorescence intensity post-centrifugation among all trials with GFP-E. coli, in comparison to the non-centrifuged samples, yielded an average magnification of 1.7 times the initial intensity.

DISCUSSION.

Due to the high prevalence of bacterial diseases worldwide and the limitations of traditional centrifugation for resource-poor communities at risk, there is a need for inexpensive technology to prepare samples for diagnosis [6]. These devices must also be easy-to-use and durable, allowing for point-of-care diagnosis in areas where sample preparation has traditionally been difficult. Additionally, granted the non-invasive and easy collection of saliva samples for bacterial detection, it is necessary to test the efficiency of these devices in preparing saliva for diagnosis. Furthermore, given the variability in viscosity among biofluids, it is important that these devices can handle more viscous biofluids, such as plasma or viscous saliva [21-22].

This study describes the design of a 3D-printed, hand-powered, and low-cost centrifuge and demonstrates its efficiency in sample compaction compared to a commercial centrifuge, in separating samples of differing viscosities, and in separating E. coli from saliva as an indicator of the centrifuge’s potential for bacterial compaction.

3D printing as a means of production proved to be an effective way to manufacture strong, reliable, and long-lasting devices that bypass previous issues with durability faced when using cardboard or paper. Using 3D printing enables these centrifuges to be mass-produced at low cost, providing resource-poor communities with durable, affordable, and efficient centrifugation [16].

The fluorescence intensity and separation percentage of the handheld device when comparison to a commercial centrifuge proved to be comparable in range to that of the commercial centrifuge set to 1,000 RPM or 63 RCF based on the lack of significant difference between the two (Figure 4). This range provides important information for determining the device’s various applications. Applications of centrifugation in this range include, but are not limited to, plasma separation and drug preparation [10, 24].

Despite the decrease in the separation of beads at increasing concentrations of glycerol, there was still a significant difference in bead separation up to the 30% glycerol dilutions. This shows the range of viscosities at which this device is helpful in compacting bacteria (Figure 5). The lack of a significant difference between 40% and 50% glycerol shows the point of diminishing return where, after 2 minutes of spinning, this centrifuge is unable to efficiently compact the sample. This data suggests a range of applicable viscosities for this centrifuge from 0.1 mPa·s (0% glycerol) to 5.23 mPa·s (30% glycerol) [25]. This range encompasses many biofluids such as blood, plasma, and most saliva and demonstrates the widescale application this device has for biofluid compaction and separation in a variety of contexts [22, 27].

Baseline testing of GFP-E. coli-spiked saliva suggested that the handheld device is an effective tool for compacting pathogenic bacteria in saliva with a significant increase in fluorescent intensity for the centrifuged saliva containing GFP-E. coli. Additionally, the lack of fluorescence and concentration of the non-spiked control group shows that the compaction observed in the GFP-E. coli spiked saliva was due to the compaction of GFP-E. coli and not of the saliva itself (Figure 6). This is important granted the application of this device in increasing the detection of pathogenic bacteria. Medical experts require an increase in bacteria concentration relative to biofluid samples which necessitates the compaction of bacteria without biofluid compaction.

When looking at data across all trials of saliva spiked with GFP-E. coli, the significant increase in fluorescence intensity of the centrifuged saliva samples when compared to the non-centrifuged samples demonstrated the significant increase in concentration of the GFP-E. coli, within the saliva samples (Figure 7). With an average of 1.729 times the initial fluorescence, the increase in final fluorescence intensity showed how the handheld centrifuge was able to concentrate pathogenic bacteria in saliva in a way that could potentially lead to increased detection during real world application. This increase in the concentration of bacteria in biofluid samples like saliva is important in medical diagnostics for fielding more reliable and confident diagnoses when looking for analytes that exist at low concentrations [5].

ACKNOWLEDGMENTS.

Thank you to Luke Whitehead and Dr. Andrea Locke for their support and mentorship, along with Dr. Pamela Popp for her advice and the entire SSMV and Locke Lab for making this research possible.

SUPPORTING INFORMATION.

Supporting information includes prototype, centrifuge dimensions, spectroscopy methods, saliva collection, E. coli growth, statistical analysis, limitations, next steps.

REFERENCES

1. P. M. Bossuyt, J. B. Reitsma, K. Linnet, K. G. Moons, Beyond Diagnostic Accuracy: The Clinical Utility of Diagnostic Tests. Clin. Chem. 58, 1636–1643 (2012).
2. P. Anker, H. Mulcahy, X. Qi Chen, M. Stroun, Detection of Circulating Tumour DNA in the Blood (Plasma/Serum) of Cancer Patients. Cancer Metastasis Rev. 18, 65–73 (1999).
3. W. G. Pitt et al., Rapid separation of bacteria from blood—review and outlook. Biotechnol. Prog. 32, 823–839 (2016).
4. K. M. Rieger-Christ et al., Identification of fibroblast growth factor receptor 3 mutations in urine sediment DNA samples complements cytology in bladder tumor detection. Cancer 98, 737–744 (2003).
5. M. S. Bhamla et al., Paperfuge: Hand-powered ultralow-cost paper centrifuge. Nat. Biomed. Eng. 1, 0009 (2017).
6. D. Mabey, R. W. Peeling, A. Ustianowski, M. D. Perkins, Diagnostics for the developing world. Nat. Rev. Microbiol. 2, 231–240 (2004).
7. P. LaBarre, D. Boyle, K. Hawkins, B. Weigl, Instrument-free nucleic acid amplification assays for global health settings. Proc. SPIE 8029, 10.1117/12.882868 (2011).
8. B. Li et al., Integrated hand-powered centrifugation and paper-based diagnosis with blood-in/answer-out capabilities. Biosens. Bioelectron. 165, 112282 (2020).
9. M. Urdea et al., Requirements for high impact diagnostics in the developing world. Nature 444, 73–79 (2006).
10. J. J. Franco, T. Nagata, T. Okamoto, S. Mukai, An ultralow-cost portable centrifuge from discarded materials for medical applications. Sci. Rep. 13, 3081 (2023).
11. Z.-R. Bi et al., Development of a Handheld Nano-centrifugal Device for Visual Virus Detection. J. Anal. Test. 6, 353–364 (2022).
12. J. Brown et al., A Hand-Powered, Portable, Low-Cost Centrifuge for Diagnosing Anemia in Low-Resource Settings. Am. J. Trop. Med. Hyg. 85, 327–332 (2011).
13. A. P. Wong, M. Gupta, S. S. Shevkoplyas, G. M. Whitesides, Egg beater as centrifuge: isolating human blood plasma from whole blood in resource-poor settings. Lab. Chip 8, 2032–2037 (2008).
14. C.-H. Liu et al., Blood Plasma Separation Using a Fidget-Spinner. Anal. Chem. 91, 1247–1253 (2019).
15. I. Michael et al., A fidget spinner for the point-of-care diagnosis of urinary tract infection. Nat. Biomed. Eng. 4, 591–600 (2020).
16. G. Byagathvalli, A. Pomerantz, S. Sinha, J. Standeven, M. S. Bhamla, A 3D-printed hand-powered centrifuge for molecular biology. PLOS Biol. 17, e3000251 (2019).
17. L. F. Hofman, Human Saliva as a Diagnostic Specimen. J. Nutr. 131, 1621S-1625S (2001).
18. Y.-H. Lee, D. T. Wong, Saliva: An emerging biofluid for early detection of diseases. Am. J. Dent. 22, 241–248 (2009).
19. J. Slots, H. Slots, Bacterial and viral pathogens in saliva: disease relationship and infectious risk. Periodontol. 2000 55, 48–69 (2011).
20. T. Takeshita et al., Bacterial diversity in saliva and oral health-related conditions: the Hisayama Study. Sci. Rep. 6, 22164 (2016).
21. W. J. Akers, M. A. Haidekker, Precision Assessment of Biofluid Viscosity Measurements Using Molecular Rotors. J. Biomech. Eng. 127, 450–454 (2005).
22. P. J. F. Rantonen, J. H. Meurman, Viscosity of whole saliva. Acta Odontol. Scand. 56, 210–214 (1998).
23. F. Momen-Heravi et al., Impact of Biofluid Viscosity on Size and Sedimentation Efficiency of the Isolated Microvesicles. Front. Physiol. 3 (2012).
24. S. Ghanaati et al., The role of centrifugation process in the preparation of therapeutic blood concentrates: Standardization of the protocols to improve reproducibility. Int. J. Growth Factors Stem Cells Dent. 2, 41 (2019).
25. J. A. Trejo González, M. P. Longinotti, H. R. Corti, The Viscosity of Glycerol−Water Mixtures Including the Supercooled Region. J. Chem. Eng. Data 56, 1397–1406 (2011).
26. G. Késmárky, P. Kenyeres, M. Rábai, K. Tóth, Plasma viscosity: a forgotten variable. Clin. Hemorheol. Microcirc. 39, 243–246 (2008).
27. Y. Tomiyama, J. Johnny E. Brian, M. M. Todd, Plasma viscosity and cerebral blood flow. Am. J. Physiol.-Heart Circ. Physiol. 279 (2000).

---

---

Posted by buchanle on Monday, May 18, 2026 in May 2026.

Tags: 3D Printing, Centrifugation, Diagnostics