An all-in-one microfluidic SlipChip for power-free and rapid biosensing of pathogenic bacteria

Abstract

Point-of-care testing of pathogens is becoming more and more important for the prevention and control of food poisoning. Herein, a power-free colorimetric biosensor was presented for rapid detection of Salmonella using a microfluidic SlipChip for fluidic control and Au@PtPd nanocatalysts for signal amplification. All the procedures, including solution mixing, immune reaction, magnetic separation, residual washing, mimicking catalysis and colorimetric detection, were integrated on this SlipChip. First, the mixture of the bacterial sample, immune magnetic nanobeads (IMBs) and immune Au@PtPd nanocatalysts (INCs), washing buffer and H₂O₂–TMB chromogenic substrate were preloaded into the sample, washing and catalysis chambers, respectively. After the top layer of this SlipChip was slid to connect the sample chamber with the separation chamber, the mixture was moved back and forth through the asymmetrical split-and-recombine micromixer by using a disposable syringe to form the IMB–Salmonella–INC sandwich conjugates. Then, the conjugates were captured in the separation chamber using a magnetic field, and the top layer was slid to connect the washing chamber with the separation chamber for washing away excessive INCs. Finally, the top layer was slid to connect the catalysis chamber with the separation chamber, and the colorless substrate was catalyzed by the INCs with peroxidase-mimic activity to generate color change, followed by using a smartphone app to collect and analyze the image to determine the bacterial concentration. This all-in-one microfluidic biosensor enabled simple detection of Salmonella as low as 101.2 CFU mL⁻¹ within 30 min and was featured with low cost, straightforward operation, and compact design.

---

1. Introduction

Infectious diseases caused by pathogens such as bacteria, viruses, fungi and parasites are increasing worldwide, leading to significant threats to public health and the global economy.^1,2 Pathogenic infections can cause diarrhea, fever, pneumonia, enteritis, sepsis, and even death. The World Health Organization estimated that 70% of foodborne illnesses were caused by pathogenic bacteria. Every year, 7.7% of people suffered from foodborne diseases, and 7.5% of global deaths resulted from foodborne illnesses.³ Screening of pathogenic bacteria, such as Salmonella, Escherichia coli O157:H7, Staphylococcus aureus, Listeria monocytogenes, etc., has become an important measure to ensure food safety, and point-of-care testing (POCT) techniques are demanded.

Existing bacterial detection methods, including cultures, PCR and ELISA, are not applicable to POCT due to either a long experimental time, sophisticated pretreatment, or unsatisfactory sensitivity. Various biosensors, including electrochemical, optical, magnetic, piezoelectric, etc.,^4–10 have been reported for fast screening of pathogenic bacteria. Besides, microfluidic chips were featured with compact size, automatic operation and unique portability, and often combined with biosensors as potential alternatives for POCT of pathogenic bacteria.^11–16 To date, many microfluidic chips relied on precise pumps and/or valves with external electrical power to automatically or semi-automatically operate the fluids, however this greatly limited their practical in-field applications. Thus, some power-free microfluidic chips were developed, such as finger-actuated,^17–19 capillary-driven^20–22 and SlipChip.^23–25 In particular, the SlipChip had drawn much attention since it was first reported by Rustem Ismagilov's group using digital RT-PCR on a rotational SlipChip for quantification of nucleic acids.²⁶ Besides, an interesting microfluidic SlipChip was exploited for POCT detection of prostate-specific antigen (PSA) with the ELISA-based method.²⁷ It was demonstrated to sensitively detect PSA as low as 1.9 ng mL⁻¹. However, the ELISA reaction processes required many steps of manual operations and thus limiting its POCT applications. In addition, micromixers often held great importance in the design of microfluidic SlipChips, as it effectively enhanced the reaction efficiency. A variety of micromixers have been exploited based on diverse geometric structures. Among them, asymmetrical split-and-recombine (ASAR) micromixers were reported and demonstrated to exhibit excellent mixing efficiency within a wide range of flow rates.^28,29 Thus, the combination of the SlipChip and ASAR micromixer might offer a promising pathway for POCT screening of pathogens.

To enhance the sensitivity of biosensors, various nanomaterials with mimetic-enzyme properties, which are also called nanocatalysts, such as Au/Ag/Pt/Pd nanoparticles, MnO₂/CoO₂/CeO₂ nanoflowers or nanoparticles, and MOFs, have attracted considerable attention and were employed to enhance colorimetric signals by virtue of mimicking catalytic ability.^30–37 Recently, much attention was paid to multifunctional hybrid nanocatalysts (such as Au@Pt, Au@Pd, Pd@Pt, Au@PtPd, MOF-Pt, MOF-PdPt) since they were featured with outstanding properties of tunable size, good biocompatibility and excellent catalytic activity.^38–42 A recent study compared the peroxidase-like catalytic activity of Au@Pt, Au@Pd, and Au@PtPd nanocatalysts. Au@PtPd showed the highest catalytic performance among these nanocatalysts and improved the sensitivity around 100 times.⁴³ Besides, colorimetric biosensors generally employed benchtop microplate readers to determine the color change, which greatly limited their POCT applications. In the past decade, smartphones were frequently reported as a good alternative for image or video processing.^44,45 Thus, the integration of a microfluidic SlipChip for power-free operation, Au@PtPd nanocatalysts for signal amplification and a smartphone for image analysis might enable sensitive POCT detection of pathogenic bacteria.

Here, a power-free microfluidic biosensor was designed for simple and rapid detection of Salmonella using a microfluidic SlipChip, a disposable syringe, and a smartphone. From Scheme 1, the microfluidic biosensor consisted of two PDMS layers and a 3D-printed holder. First, when the sample chamber, separation chamber, washing chamber and catalysis chamber of this chip were independently disconnected, the bacterial sample with the immune magnetic nanobeads (IMBs) and immune Au@PtPd nanocatalysts (INCs), washing buffer and H₂O₂–TMB chromogenic substrate were preloaded into the sample, washing chamber and catalysis chamber, respectively. Then, the top layer was slid to connect the sample chamber with the separation chamber, and the mixture of the bacterial sample with IMBs and INCs was moved back and forth through the ASAR micromixer by using a disposable syringe to form the IMB–Salmonella–INC sandwich conjugates, which were captured in the separation chamber by an external magnetic field. After the top layer was slid to connect the washing chamber with the separation chamber for rinsing the conjugates to remove excessive INCs, the top layer was slid to connect the catalysis chamber with the separation chamber. Finally, the colorless H₂O₂–TMB chromogenic substrate was catalyzed by the INCs on the sandwich conjugates to generate color change, followed by image acquisition and processing to determine the target bacteria using a smartphone app. Besides, the result was combined with the sampling time and location and transmitted wirelessly to a cloud platform for risk assessment and early warning.

Scheme 1 Schematic of the microfluidic SlipChip for biosensing of Salmonella. (A) 3D design of the microfluidic SlipChip. (B) Structure of the microfluidic SlipChip. (C) Mechanism of the colorimetric biosensor. (D) Operation of the microfluidic SlipChip for Salmonella detection.

2. Materials and methods

2.1 Materials

The detailed information for the reagents and materials were provided in the ESI.†

2.2 Culture of the bacteria

The detailed procedure for the culture of bacterial strains was provided in the ESI.†

2.3 Preparation of the spiked bacterial samples

The detailed preparation of the spiked bacterial samples was provided in the ESI.†

2.4 Design and fabrication of the SlipChip

This microfluidic SlipChip was designed to perform the whole bacterial detection procedure, including sample loading, solution mixing, immune reaction, magnetic separation, residual washing, mimicking catalysis and colorimetric detection. As shown in Scheme 1B and Fig. S1,† the microfluidic SlipChip consists of two PDMS layers with the chambers and channels, and a 3D printing holder with a magnet (material: NdFeB, grade: N52) for preventing the deformation of the top PDMS layer during sliding and performing the magnetic separation of the IMBs and their conjugates.

As shown in Fig. 1A, the PDMS top layer includes a sample chamber (∼150 μL) for loading the mixture of IMBs, INCs and the bacterial sample, a passive ASAR micromixer for mixing the mixture, an incubation channel for performing the immunoreaction, a washing chamber (∼200 μL) for preloading the wash buffer and a catalysis chamber (∼50 μL) for preloading the substrate. The bottom layer contains a shallow separation chamber for capturing the conjugates and some microfluidic channels for connecting the chambers. The sample, separation, washing and catalysis chambers were initially disconnected, and the mixture, wash buffer and substrate were preloaded into their respective chambers through their corresponding inlets prior to bacterial detection.

Fig. 1 (A) Structure of the SlipChip. (B) Fabrication process of the SlipChip. (C) Cross-sectional image of two PDMS layers without and with the silicon oil layer. (D) Evaluation of the SlipChips with different widths and same depth of channels (D₁ and D₄: 0.5 mm wide and 1 mm deep; D₂ and D₅: 1 mm wide and 1 mm deep; D₃ and D₆: 2 mm wide and 1 mm deep).

For sliding of this microfluidic SlipChip, a thin silicon oil layer was spin-coated between these two PDMS layers for their smooth sliding based on our previous report with some modifications.⁴⁶ As shown in Fig. 1B, the silicon oil was first spun on a glass wafer at 2000 rpm to form a silicon oil film. Then, the top PDMS layer was placed on the oil film to sufficiently and evenly coat with a thin layer of oil. After the layer was peeled off, it was placed onto the bottom layer in the holder, followed by waiting for 10 min to form a uniform silicon oil layer.

For fabrication of this SlipChip, the molds of the PDMS layers were first drawn using the Solidworks software and fabricated using a Objet30 Pro 3D printer. Then, the PDMS prepolymer and curing agent were prepared at a mass ratio of 10 : 1 to cast the molds. After vacuuming for 30 min to remove the air, they were cured at 65 °C for 6–8 h and peeled off from the mold. Finally, these two PDMS layers were assembled through the silicon oil layer to form the SlipChip.

2.5 Synthesis of the Au@PtPd NCs

The Au@PtPd NCs were synthesized according to the reported protocol with some modifications.⁴³ Briefly, 0.75 mL trisodium citrate (34 mM, 1 wt%) and 0.25 mL HAuCl₄·3H₂O (25 mM, 1 wt%) was first added into a clean glass vial and mixed at room temperature. Then, 0.25 mL deionized water was added and mixed. After incubation for 15 min, 3.125 mL H₂PtCl₆·6H₂O (5 mM) and 3.125 mL Cl₄Na₂Pd (5 mM) were added and mixed for 30 min, followed by adding 1 mL ascorbic acid (0.25 M) and mixing for 3 h with vigorous stirring. Finally, the mixture was centrifuged at 5630 × g for 10 min and washed twice with acetone and three times with water to remove the residuals, and the Au@PtPd NCs were obtained and stored at 4 °C.

2.6 Preparation of the IMBs and INCs

The IMBs were prepared through covalent binding to modify the polyclonal antibodies onto the carboxylated MBs, and the detailed procedure could be found in our previous work.⁴⁷ The INCs were prepared through electrostatic adsorption to conjugate the monoclonal antibodies with the Au@PtPd NCs. First, 0.5 mL Au@PtPd NCs were centrifuged at 5630 × g and resuspended in 0.5 mL PB (10 mM, pH 7.4). Then, 6 μL anti-Salmonella monoclonal antibodies (2.5 mg mL⁻¹) were added and incubated for 3.5 h. After 60 μL 10% BSA was added and incubated for 1.5 h to block the excessive binding sites, the mixture was centrifuged at 4450 × g for 10 min and washed three times with 1% BSA to remove the unbound antibodies. Finally, the precipitate was resuspended in 150 μL PBS containing 1% BSA and stored at 4 °C.

2.7 Establishment of the calibration curve

The establishment of the calibration curve for this biosensor was the basis for the determination of target bacteria in unknown samples. Thus, pure target bacterial cultures at different concentrations from 2.0 × 10¹ to 2.0 × 10⁶ CFU mL⁻¹ were prepared and detected by this biosensor to build the calibration curve. As illustrated in Fig. 1D, before a 150 μL mixture containing the IMBs, INCs and bacterial sample was loaded into the sample chamber, 200 μL wash buffer (PBS) and 50 μL substrate (H₂O₂–TMB) were preloaded into their corresponding chambers, respectively. First, the top PDMS layer of the SlipChip was slid to connect the sample chamber with the separation chamber, followed by moving the mixture back and forth through the ASAR micromixer and incubation channel by using a disposable syringe to form the IMB–Salmonella–INC sandwich conjugates. Then, the conjugates were captured in the separation chamber by the magnet in the holder while the sample background flowed into the waste chamber. After the top layer was slid to connect the washing chamber with the separation chamber for washing the conjugates to remove the excessive INCs, the top layer was finally slid to connect the catalysis chamber with the separation chamber and the colorless H₂O₂–TMB substrate was catalyzed by the INCs on the sandwich conjugates to generate color change, followed by collecting and analyzing the image of the catalysate using a smartphone app to obtain its saturation. The calibration curve of this colorimetric biosensor was constructed based on the linear correlation between the saturation of the catalysate and the logarithm of bacterial concentration. The entire procedure is also demonstrated in Video S3.†

3. Results and discussion

3.1 Evaluation of the microfluidic SlipChip

In this microfluidic SlipChip, the silicone oil was used to separate the two PDMS layer and smoothen their sliding to achieve the entire bacterial detection procedure step by step. The thickness of the oil layer and the width of the channels were important for smooth sliding without leakage, and thus were observed using a microscope. From the cross-sectional images in Fig. 1C, the thickness of the oil layer was ∼20 μm. To check the tight connection of two PDMS layers with the channels, two PMDS layers with different widths (0.5, 1 and 2 mm) and the same depth (1 mm) of the straight channels and a diamond chamber were fabricated and separated by the oil layer to check the smoothness and leakage. The blue dye was pipetted into the chamber, and a syringe was used to push the dye for checking the connection. From Video S1† and Fig. 1D, when the width was 0.5 mm, it was hard for the dye to flow through the connected channels. This might be because the channel was too narrow and had a high surface tension. While the width increased to 1 mm and 2 mm, the dye could easily flow back and forth through the connected channels and no leakage was observed. Furthermore, it was found that silicone oil was more likely to penetrate the channels with a width of 2 mm and result in the undesired blockage. Therefore, the width and depth of 1 mm was used for the channels.

To further assess the feasibility of this SlipChip, several dyes were first loaded into the chambers. Then, the chip was slid to connect different channels in turn and the dyes were moved back and forth using a syringe. From Video S2,† the dyes flowed from their chambers to next chambers smoothly when their channels are connected, indicating the feasibility of this SlipChip.

3.2 Performance of the ASAR micromixer

The conjugation of the target bacteria with the IMBs and INCs mainly depended on the mixing effect of the ASAR micromixer. Thus, the finite element analysis software COMSOL was utilized to simulate the ASAR micromixer for assessing its mixing efficiency. The detailed parameters for this micromixer were shown in Fig. S4.† From Fig. 2A, this micromixer shows an excellent mixing efficiency of ∼90%. To assess the actual mixing effect of this micromixer, deionized water and red dye were first injected into a chamber successively (Fig. S3†), and then moved back and forth through the ASAR micromixer using a syringe. Finally, the images were collected using the smartphone app after different times of movement (Fig. 2B), and processed using ImageJ software to obtain their respective standard deviations. The mixing efficiency (ME_i) was calculated by the equation:where, SD_i is the standard deviation after the ith back-and-forth movement and SD₀ is the standard deviation before the back-and-forth movement. As illustrated in Fig. 2C, the mixing efficiency increased with the number of the back-and-forth movement, and achieved over 90% after the 7th movement. Besides, the time for mixing was also important for the conjugation of target bacteria with the IMBs and INCs. Thus, different times (2, 6, 10 and 15 min) were first applied for mixing the IMBs and target bacteria at a concentration of 5.0 × 10³ CFU mL⁻¹ using this micromixer, respectively. Then, the conjugates were captured in the separation chamber using the magnet. After the solution was washed away, the magnet was removed and the captured bacteria were flushed out using PBS buffer, which were finally culture plated for enumeration. The capture efficiency was calculated as the ratio of the captured bacteria to the original one. From Fig. 2D, the capture efficiency reaches 95% when the time is 10 min or longer, indicating that 10 min is enough for specific capture of target bacteria.

Fig. 2 (A) COMSOL simulation of the ASAR micromixer. (B) Experimental mixing result of red dye–water mixture with the number of the back-and-forth movement from the ASAR micromixer. (C) Mixing efficiency vs. the number of the back-and-forth movement (N = 3). (D) Optimization of the time of back and forth for capturing target bacteria (N = 3). (E) Comparison of the capture efficiency for the target bacteria of dynamic mixing with/without the ASAR micromixer and static incubation (N = 3). (F) Capture efficiency for different flow rates (N = 3).

Compared to other micromixers, this ASAR micromixer showed a high mixing efficiency because both the asymmetrical structure and the split-and-recombine structure could enhance the mixing of the mixture. Besides, the mixture was moved the mixture back and forth from the diamond chamber to the mixing and incubation channels, which was verified with a good mixing effect as well in our previous studies.⁴⁸ To further access the performance of the micromixer, the IMBs and the bacterial sample (5.0 × 10³ CFU mL⁻¹) were loaded into the sample chamber, and three different strategies for capturing the target bacteria were compared: (1) dynamic mixing with this ASAR micromixer; (2) dynamic mixing without this ASAR micromixer; (3) static incubation. From Fig. 2E, the dynamic mixing strategy with the ASAR micromixer has a capture efficiency of 95%, which is significantly higher than the one without the micromixer (71%) and the static incubation strategy (36%). This highest capture efficiency for the dynamic mixing strategy with the ASAR micromixer was attributed to the excellent mixing effect of the ASAR micromixer and the continuous contact of target bacteria with the IMBs in a back-and-forth flow.

In this study, the disposable syringe was employed as the alternative for electric-controlled micropumps to manually move the mixture back and forth, and the flow rate might be not easily controlled. To check the impact of the flow rate, a precise peristaltic pump was used to provide four different flow rates for mixing the IMBs and bacterial sample. From Fig. 2F, when the flow rate increases from 0.5 mL min⁻¹ to 1.25 mL min⁻¹, the capture efficiency basically remains at the range of 85–95%, indicating that the flow rate has no obvious impact on the capture of target bacteria. This was attributed to the good dynamic mixing effect of the ASAR micromixer to make the target bacteria have a continuous contact with the IMBs at any flow rate to form the similar amount of IMB–bacteria conjugates.

3.3 Characterization of the Au@PtPd NCs

The Au@PtPd NCs were used as signal probe to label the target bacteria and are important for improving the sensitivity of the biosensor. Thus, the characterization of the nanocatalysts was conducted. The morphology of the Au@PtPd NCs was characterized by scanning electron microscopy (SEM), transmission electron microscopy (TEM), high-resolution TEM (HRTEM), and high-angle annular dark field (HAADF). From Fig. 3(A–D), Au@PtPd NCs show a uniform structure with an average size of ∼50 nm, which is basically consistent with the dynamic light scatting (DLS) result (Fig. 3I). Besides, the energy dispersive spectrometer (EDS) spectra are shown in Fig. 3(E–H) and S2† to verify the elementary composition of Au (yellow), Pt (red) and Pd (green), indicating the successful synthesis of the Au@PtPd NCs. The modification of the antibodies onto the Au@PtPd NCs was vital for specifically labelling the target bacteria. Thus, DLS and Zeta potential analysis were conducted to characterize the Au@PtPd NCs and immune Au@PtPd NCs. Fig. 3I shows that the average size of the immune Au@PtPd NCs (∼60 nm) is ∼10 nm larger than that of the Au@PtPd NCs (∼50 nm) due to the successful modification of the anti-Salmonella antibodies. Fig. 3J shows that the zeta potential of the Au@PtPd NCs was −46 mV and changed to −33 mV after the antibodies were modified onto the nanocatalysts. This also demonstrated the successful modification of the Au@PtPd NCs with the antibodies.

Fig. 3 Characterization of the Au@PtPd NCs. (A) The SEM image. (B) The TEM image. (C) The HRTEM image. (D) The HAADF image. (E–H) The EDS spectra. (I) The DLS results for the Au@PtPd NCs and immune Au@PtPd NCs. (J) The zeta potential of the Au@PtPd NCs and immune Au@PtPd NCs (N = 3).

3.4 Concept proof of the biosensor

The biosensor was based on the mimicking catalysis of H₂O₂–TMB by INCs on the sandwich conjugates to generate color change. To validate this concept, the INCs at different concentrations from 0 to 100 μg mL⁻¹ were employed to catalyze H₂O₂–TMB, and the resulting image for each catalysate was collected and processed by the self-developed smartphone app to get its saturation. As shown in Fig. 4A, the blue color changes from light to dark and the saturation grows with the INC concentration. As shown in Fig. 4B, a linear relationship is found between the saturation and the INC concentration (1.56–50 μg mL⁻¹), proving the feasibility of this concept.

Fig. 4 (A) The saturation for different concentrations of INCs from 0 to 100 μg mL⁻¹ to catalyze H₂O₂–TMB (N = 3). (B) The linear relationship between the catalysate saturation and the INC concentration from 1.56 to 50 μg mL⁻¹ (N = 3). (C) The optimization of the amount of IMBs and INCs. (D) The optimization of the catalysis time (N = 3).

In this biosensor, the IMB–bacteria–INC conjugates were formed through a one-step immune reaction, and the IMBs might be competitive with the INCs in their reaction with target bacteria. Thus, the amounts for the IMBs and INCs were optimized through the orthogonal experiments, i.e., different amounts of the IMBs from 10 to 25 μg and the INCs from 5 to 20 μg were used to react with Salmonella typhimurium at 3.0 × 10⁵ CFU mL⁻¹ simultaneously. As shown in Fig. 4C, the saturation of the resulting catalysate gradually increases when the INC amount changes from 5 to 15 μg, and more amount (20 μg) does not lead to the saturation increase, indicating that 15 μg of the INCs is sufficient for labeling Salmonella. Besides, the saturation of the catalysate gradually increases as the IMBs increase from 10 to 20 μg. However, when the amount of IMBs further increase to 25 μg, a decrease in saturation is observed. This could be attributed to the competition between IMBs and INCs for the immune reaction with the target bacteria. Thus, 20 μg IMBs and 15 μg INCs were selected for this biosensor. In addition, the catalysis time from 1 to 9 min was also optimized using the INCs to catalyze the chromogenic substrate. From Fig. 4D, the saturation changes from 74 to 210 as the catalysis time increases from 1 to 5 min, and remains basically at the same level when the catalysis time increases to 7 min, indicating that 5 min is sufficient for the INCs to catalyze the H₂O₂–TMB substrate.

3.5 Evaluation of the biosensor

To evaluate the sensitivity of this biosensor, three parallel tests on Salmonella typhimurium at 2.0 × 10¹–2.0 × 10⁶ CFU mL⁻¹ were carried out using this biosensor under the optimal conditions. From Fig. 5A, the saturation of the catalysate gradually increases with the concentration of bacteria. From Fig. 5B, the linear correlation is observed between the saturation (S) and the logarithm of the bacterial concentration (C), which could be represented as S = 25.375 × ln(C) − 54.17 (R² = 0.98). The limit of detection (LOD) was calculated to be 101 CFU mL⁻¹ based on the 3× signal-to-noise ratio.⁴⁹ Besides, TEM imaging was employed to further confirm the formation of the IMB–bacteria–INC sandwich conjugates (Fig. 5C).

Fig. 5 (A) The saturation of the catalysate for different concentrations of Salmonella. (B) The linear correlation between the saturation (S) and the logarithm of bacterial concentration (C) from 2.0 × 10² to 2.0 × 10⁵ CFU mL⁻¹ (N = 3). (C) The TEM image of the IMB–Salmonella–INC conjugates (the inset was the TEM image of IMBs). (D) The specificity of this biosensor (N = 3). (E) The difference between different operators of this biosensor. (F) Radar map for comparison of this biosensor with conventional ELISA.

The specificity of this biosensor was verified through detecting the target bacteria (Salmonella typhimurium) at a concentration of ∼10⁴ CFU mL⁻¹, the non-target bacteria (Staphylococcus aureus, Bacillus cereus, E. coli O157:H7, Listeria monocytogenes) at a concentration of ∼10⁶ CFU mL⁻¹ and their mixture. The polyclonal antibodies against Salmonella typhimurium modified on the surface of the MBs and the monoclonal antibodies against Salmonella typhimurium modified on the surface of the NCs could specifically react with Salmonella typhimurium to form the IMB–Salmonella–INC sandwich conjugates. From Fig. 4D, the saturation levels for the target bacteria (211.3) and the mixture (216.9) are obviously higher than those for the negative control (51.3) and the non-target bacteria (49.9 for Staphylococcus, 46.8 for Bacillus, 64.3 for E. coli, 56.9 for Listeria), illustrating the satisfactory specificity of this biosensor. Besides, to check the difference between different operators, target bacteria at the same concentration (3.0 × 10³ CFU mL⁻¹) was respectively detected using this biosensor by three operators after they were simply trained. From Fig. 5E, the saturation levels are basically similar, demonstrating that this biosensor could be operated by different operators.

Compared with the conventional ELISA method and some recent reported methods, this microfluidic biosensor showed some merits (Fig. 5F, Tables S1 and S2†). First, this microfluidic SlipChip integrated the entire bacterial detection procedure onto one single chip, avoiding the requirement of professional laboratories and well-trained technicians. Meanwhile, this biosensor achieved the collection and analysis of colorimetric signals through the simple smartphone app, avoiding the need for bulky instruments. Furthermore, the Au@PtPd NCs amplified the detection signals, enhancing the sensitivity of bacterial detection. Notably, this biosensor achieved the bacterial detection in 30 min at the cost of ∼$8.9, enhancing the possibility of POCT applications. This sensor exhibited high integration, short experimental time, good portability and comparable sensitivity, and showed potential for POCT of pathogenic bacteria.

3.6 Practicability of this biosensor

The practical applicability of the biosensor was assessed by detecting Salmonella in spiked milk and chicken samples at concentrations of 3.0 × 10²–3.0 × 10⁴ CFU mL⁻¹. The spiked samples were parallel detected using this biosensor and culture plating, and the recovery was calculated as the ratio of the result from this biosensor to that from culture plating. As shown in Table 1, the recovery for the milk samples ranges from 82% to 115% with the mean relative standard deviation of less than 11%, and the recovery for the spiked chicken samples ranges from 92% to 110% with the mean relative standard deviation of less than 12%. There was a slight difference between the biosensor and the culture plating, which might be attributed to the influence of the background in the milk and chicken samples, such as fats, proteins and other molecules. These results indicated that this biosensor exhibited satisfactory applicability to detect bacteria in the spiked food samples. Besides, the proposed biosensor might be used for in-field screening of Salmonella in poultry or milk supply chains for early warning of potential Salmonella contaminations, and could be extended for the detection of other foodborne pathogens and biomarkers by changing their specific antibodies.

Table 1 Detection of Salmonella in spiked chicken and milk samples using this biosensor (N = 3)

Spiked (CFU mL^−1)	Milk			Chicken
	Detected (CFU mL^−1)	Recovery (%)	RSD (%)	Detected (CFU mL^−1)	Recovery (%)	RSD (%)
3.0 × 10^2	245.8	81.95	10.97	275.3	91.78	11.50
3.0 × 10^3	2755.0	91.83	9.81	3260.7	108.69	6.97
3.0 × 10^4	31 121.1	103.74	6.61	32 981.8	109.94	7.06

---

Conclusions

Herein, a power-free colorimetric biosensor was successfully developed using the microfluidic SlipChip and Au@PtPd nanocatalysts for POCT of Salmonella. This microfluidic SlipChip was proved to perform the entire bacterial detection procedure from mixing, separation, washing, catalysis and detection. This ASAR micromixer demonstrated a good mixing effect and the Au@PtPd nanocatalysts were verified to enhance the sensitivity of this biosensor. The all-in-one biosensor was able to detect Salmonella from 2.0 × 10² to 2.0 × 10⁵ CFU mL⁻¹ in 30 min with the detection limit of 101.2 CFU mL⁻¹, and had the potential for POCT of pathogenic bacteria due to its low cost, straightforward operation, compact design and high integration.

Data availability

The data supporting this article have been included in the main text or the ESI.† The code of the smartphone app is also available in the ESI.†

Conflicts of interest

The authors have no conflicts of interest to declare that are relevant to the content of this article.

Acknowledgements

This study was supported by the National Natural Science Foundation of China (Grant no. 32071899).

References

1. X. Cheng, C. C. Wang, A. Liu, Y. Liu, J. Dong, Z. Wang, W. Wei and S. Liu, Spectrochim. Acta, Part A, 2023, 285, 121830.
2. Y. He, Q. Hu, S. San, T. Kasputis, M. G. D. Splinter, K. Yin and J. Chen, TrAC, Trends Anal. Chem., 2023, 168, 117342.
3. H. Lee and Y. Yoon, Food Sci. Anim. Resour., 2021, 41, 1–7.
4. S. Wang, X. Zhu, Q. Meng, P. Zheng, J. Zhang, Z. He and H. Jiang, Biosens. Bioelectron., 2021, 183, 113186.
5. S. Wang, C. Sun, Q. Hu, S. Li, C. Wang, P. Wang and L. Zhou, Biochem. Eng. J., 2020, 156, 107513.
6. A. B. Pebdeni, A. Roshani, E. Mirsadoughi, S. Behzadifar and M. Hosseini, Food Control, 2022, 108822,  DOI:10.1016/j.foodcont.2022.108822.
7. S. Rodphukdeekul, M. Tabata, C. Ratanaporncharoen, Y. Takeuchi, P. Somboon, W. Boonlue, Y. Miyahara and M. Sriyudthsak, Chemosensors, 2021, 9, 159.
8. L. Wei, Z. Wang, C. Feng, Y. Xianyu and Y. Chen, Anal. Chem., 2021, 93, 6613–6619.
9. Y. Xianyu, Y. Dong, Z. Wang, Z. Xu, R. Huang and Y. Chen, ACS Sens., 2019, 4, 1942–1949.
10. M. B. Kulkarni, N. H. Ayachit and T. M. Aminabhavi, Biosensors, 2023, 13, 246.
11. D. Gao, Z. Ma and Y. Jiang, TrAC, Trends Anal. Chem., 2022, 157, 116788.
12. L. M. Schmidt-Speicher and K. Länge, Curr. Opin. Electrochem., 2021, 29, 100755.
13. X. Fu, J. Sun, R. Liang, H. Guo, L. Wang and X. Sun, Trends Food Sci. Technol., 2021, 116, 115–129.
14. Y. Liu, H. Shen, X. Yang, S. Kang, L. Cai, T. Tian, R. Su, C. Yang and Z. Zhu, TrAC, Trends Anal. Chem., 2023, 158, 116894.
15. P.-L. Lonchamps, Y. He, K. Wang and X. Lu, J. Agric. Food Res., 2022, 100430,  DOI:10.1016/j.jafr.2022.100430.
16. M. B. Kulkarni, N. H. Ayachit and T. M. Aminabhavi, Biosensors, 2023, 13, 543.
17. W. Qi, S. Wang, L. Wang, X. Xi, S. Wu, Y. Li, M. Liao and J. Lin, Chin. Chem. Lett., 2023, 34, 107360.
18. N. Jin, L. Xue, Y. Ding, Y. Liu, F. Jiang, M. Liao, Y. Li and J. Lin, Biosens. Bioelectron., 2023, 220, 114844.
19. Z. Wang, Y. Wang, L. Lin, T. Wu, Z. Zhao, B. Ying and L. Chang, Biosens. Bioelectron., 2022, 195, 113663.
20. C. Achille, C. Parra-Cabrera, R. Dochy, H. Ordutowski, A. Piovesan, P. Piron, L. Van Looy, S. Kushwaha, D. Reynaerts, P. Verboven, B. Nicolai, J. Lammertyn, D. Spasic and R. Ameloot, Adv. Mater., 2021, 33, e2008712.
21. M. Yafia, O. Ymbern, A. O. Olanrewaju, A. Parandakh, A. Sohrabi Kashani, J. Renault, Z. Jin, G. Kim, A. Ng and D. Juncker, Nature, 2022, 605, 464–469.
22. M. B. Kulkarni, N. H. Ayachit, T. M. Aminabhavi and B. W. Pogue, Biochem. Eng. J., 2023, 198, 109027.
23. Z. Yu, L. Xu, W. Lyu and F. Shen, Lab Chip, 2022, 2954–2961,  10.1039/d2lc00284a.
24. Z. Yu, W. Lyu, M. Yu, Q. Wang, H. Qu, R. F. Ismagilov, X. Han, D. Lai and F. Shen, Biosens. Bioelectron., 2020, 155, 112107.
25. Y. Li, J. Xuan, Y. Song, W. Qi, B. He, P. Wang and L. Qin, ACS Nano, 2015, 10, 1640–1647.
26. F. Shen, B. Sun, J. E. Kreutz, E. K. Davydova, W. Du, P. L. Reddy, L. J. Joseph and R. F. Ismagilov, J. Am. Chem. Soc., 2011, 133, 17705–17712.
27. G. Fu, R. Hou, X. Mou and X. Li, Anal. Chem., 2021, 93, 15105–15114.
28. W. Raza, S. B. Ma and K. Y. Kim, Micromachines, 2019, 10, 711.
29. D. Jiao, R. Zhang, H. Zhang, S. Ren, H. Feng and H. Chang, J. Micromech. Microeng., 2022, 32, 075007.
30. Y. Huang, J. Ren and X. Qu, Chem. Rev., 2019, 119, 4357–4412.
31. J. Liu, L. Meng, Z. Fei, P. J. Dyson and L. Zhang, Biosens. Bioelectron., 2018, 121, 159–165.
32. R. Das, A. Dhiman, A. Kapil, V. Bansal and T. K. Sharma, Anal. Bioanal. Chem., 2019, 411, 1229–1238.
33. G. Chen, M. Jin, J. Ma, M. Yan, X. Cui, Y. Wang, X. Zhang, H. Li, W. Zheng, Y. Zhang, A. M. Abd El-Aty, A. Hacimuftuoglu and J. Wang, J. Agric. Food Chem., 2020, 68, 660–668.
34. S. B. He, F. Q. Chen, L. F. Xiu, H. P. Peng, H. H. Deng, A. L. Liu, W. Chen and G. L. Hong, Anal. Bioanal. Chem., 2020, 412, 499–506.
35. N. Cheng, C. Zhu, Y. Wang, D. Du, M.-J. Zhu, Y. Luo, W. Xu and Y. Lin, J. Anal. Test., 2018, 3, 99–106.
36. J. Wu, W. Lv, Q. Yang, H. Li and F. Li, Biosens. Bioelectron., 2021, 171, 112707.
37. M. B. Kulkarni, N. H. Ayachit and T. M. Aminabhavi, Biosensors, 2022, 12, 892.
38. O. Adeniyi, S. Sicwetsha and P. Mashazi, ACS Appl. Mater. Interfaces, 2020, 12, 1973–1987.
39. S. K. Gurmessa, L. T. Tufa, J. Kim, K.-I. Lee, Y.-M. Kim, V. T. Tran, H.-Q. Nguyen, T. S. Shim, J. Kim, T. J. Park, J. Lee and H.-J. Kim, ACS Appl. Nano Mater., 2020, 4, 539–549.
40. A. Cao, Y. Sun, F. Pei, X. Mu, B. Du, Z. Tong, Q. Hao, M. Xia, W. Lei and B. Liu, Microchem. J., 2023, 194, 109263.
41. L. Zheng, G. Cai, W. Qi, S. Wang, M. Wang and J. Lin, ACS Sens., 2020, 5, 65–72.
42. X. Wang, M. Zhang, X. Pang, K. Huang, Z. Yao, X. Mei and N. Cheng, Talanta, 2022, 247, 123503.
43. X. Ruan, V. Hulubei, Y. Wang, Q. Shi, N. Cheng, L. Wang, Z. Lyu, W. C. Davis, J. N. Smith, Y. Lin and D. Du, Biosens. Bioelectron., 2022, 208, 114190.
44. Y. Ren, J. Wei, Y. He, Y. Wang, M. Bai, C. Zhang, L. Luo, J. Wang and Y. Wang, Food Chem., 2021, 343, 128518.
45. E. M. Khalaf, H. Sanaan Jabbar, R. Mireya Romero-Parra, G. Raheem Lateef Al-Awsi, H. Setia Budi, A. S. Altamimi, M. Abdulfadhil Gatea, K. T. Falih, K. Singh and K. A. Alkhuzai, Microchem. J., 2023, 190, 108692.
46. S. Wang, G. Cai, H. Duan, W. Qi and J. Lin, Lab Chip, 2021, 22, 80–89.
47. L. Xue, R. Guo, N. Jin, S. Wang, H. Duan, W. Qi, L. Wang, Y. Zheng, Y. Li and J. Lin, Food Control, 2023, 143, 109316.
48. R. Guo, L. Xue, N. Jin, H. Duan, M. Li and J. Lin, Biosens. Bioelectron., 2022, 213, 114458.
49. S. Shan, D. Liu, Q. Guo, S. Wu, R. Chen, K. Luo, L. Hu, Y. Xiong and W. Lai, J. Dairy Sci., 2016, 99, 7025–7032.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4lc00366g

---