Non-perpendicular incidence angle of cold atmospheric-pressure plasma treatment for fabrication of silver-based surface plasmon resonance biosensor for detection of Escherichia coli in water

In this study, for the first time, to our knowledge, a biosensor was produced using cold atmospheric plasma (CAP) treatment of silver thin film surface with non-perpendicular incidence angle for identification of Escherichia coli in the distilled water. Field emission scanning electron microscopy (FESEM) exhibits that before deposition, non-perpendicular CAP treatment of glass surface substrate leads to the production of pinhole-free silver thin film. The results of atomic force microscopy (AFM), and curve fitting show that non-perpendicular CAP treatment of this pinhole-free silver thin film indicates to the appearance of Ag NPs with smaller size and larger surface area compared to untreated silver film deposited on the untreated glass substrate. The silver-based pinhole-free SPR biosensor produced with non-perpendicular CAP treatment of both glass substrate and silver film shows E. coli detection in the distilled water in the range of 10 Colony forming unit (CFU/ml) to 10 CFU/ml with better sensitivity compared to the untreated silver-based SPR biosensor.

GRAPHICAL ABSTRACT INTRODUCTION utilized to control silver thin film surface qualities such as roughness or heterogeneity for enhancing bacterial protein adsorption.
It has been reported that the perpendicular CAP treatment can improve Ag NPs interaction with gram-negative bacteria cells by decreasing nanoparticle (NP) size (Manaloto et al. 2020) and enhancing the surface hydrophilicity (Zendehnam et al. 2018). Furthermore, it has been proved that the non-perpendicular CAP treatment of silver thin film result in the significant variations in the Ag NPs size (Hosseinpour et al. 2019). In addition, the CAP treatment of the glass substrate improves the Ag NPs deposition rate, increases the thin film adhesion as well as the density and compactness of the deposited Ag NPs; leading to the fabrication of pinhole-free thin films (Lisco et al. 2017); resulting in the bacterial detection based on protein adsorption (Wigginton et al. 2010). CAP treatment is an environmentally benign technology, considered as a surface modification technique for such materials as polymers (Reche et al. 2016), or inorganic materials (e.g. metals) (Kawase et al. 2014;Hosseinpour et al. 2019). Hence, in this study, the positive effect of non-perpendicular CAP treatment on the surface of silver thin film and its glass substrate was used for the biosensor production and the SPR signals were measured using the intensity-wavelength interrogation methodology with the help of a spectrophotometer.
To investigate the influence of the non-perpendicular CAP treatment on the device performance, here, the sensitivity of non-perpendicular CAP-treated silver thin film deposited on the non-perpendicular CAP-treated glass substrate was compared to the sensitivity of untreated silver thin film deposited on the untreated glass for identification of E. coli bacteria in the distilled water.
One critical challenge for SPR measurements is to correctly build a SPR sensing platform. There are four types of methodology to measure the SPR signal. The intensity modulation is the first one, in which, the principle is to detect the intensity variation of the reflectivity in particular angle or wavelength (Prabowo et al. 2018). In this methodology as a spectrophotometer is utilized, an applicable sensor chip holder is necessary to suspend correctly the chip in the standard spectrophotometer flow cells containing the sample solutions with various concentrations for recording the reflectivity spectra. Wang et al. utilized a complicated flow cell to measure SPR signals (Wang et al. 2017). However, in this work, a simple and low-cost chip holder was designed using aided design computer-aided design (CAD) model and fabricated by the use of 3D printing technology (Shahrubudin et al. 2019) to facilitate the SPR measurement with standard spectrophotometer flow cells. This sensor holder has the advantage of reusable for a number of SPR measurements since it can be properly cleaned by the use of such disinfectant liquids as bleach.

APPJ-set up
To produce sensor chip using CAP treatment, a schematic representation of the used experimental set-up is shown in Figure 1. APPJ consisted of a dielectric, powered, and grounded electrodes and a 6 KHz high voltage pulsed DC power supply. A copper rod (length ¼ 30 mm, diameter ¼ 1 mm) and a thin copper cylindrical tube (length ¼ 15 mm) used as powered and grounded electrodes, respectively. A Pyrex tube (length ¼ 150 mm, OD ¼ 6 mm, ID ¼ 4 mm) employed as the dielectric barrier (the jet nozzle). The powered electrode was inserted in the Pyrex tube from one end, while the other tube end was surrounded by the grounded electrode. The distance between the nozzle tip and the lower edge of the grounded electrode was 5 mm. Moreover, argon gas with a purity of 99.999% used as feeding gas with flow rate of 2slm. CAP applications are mostly based on the generation capability of sufficient amounts of various reactive species (Rezaeinezhad et al. 2019). So, optical emission spectroscopy (OES) was used to analyze the existence, and intensity of any species. For this purpose, an Ocean Optic HR 2000 spectrometer was employed with the spectral range of 200-1,100 nm and optical resolution of 0.5 nm. The plasma optical emissions recorded at distance of 1 cm from plasma stream. According to the atomic spectra database lines, the recorded spectrum was analyzed and different species were determined.
Sputtering silver film on the CAP-treated and untreated glass substrates Silver thin film was prepared by DC magnetron sputtering (Hind High Vacuum, H.H.V, 12″MSPT) of silver onto untreated and CAP-treated square glass substrates (1 Â 1 cm 2 ). It has been reported that based on such SPR signal analyses as Q-factor and full width half maximum (FWHM), a thickness of 45 nm or 55 nm silver layer can generate excellent amount of surface plasmon polaritons (SPP) due to its low FWHM and high Q-factor values (Muhktar et al. 2018). Hence, the thicknesses of thin films were 45 nm with deposition rate of 1.5 nm°/S. The working condition of the magnetron sputter was 195 W with voltage of 325 V, Ag target (99.9% purity), substrate to target distance 12 cm with argon as working gas (purity 99.99%) and pressure of 3 Â 10 À2 mbar (base pressure 10 À5 mbar). The coating rate and silver film thickness were measured using a vibrating quartz crystal thickness monitor.

CAP treatment of the glass substrate
It has been reported that the deposition techniques such as sputtering, electrodeposition, and spin coating require surface treatment of the substrate with atmospheric pressure plasma prior to the deposition of the thin film (Lisco et al. 2017). This is due to the fact that atmospheric-pressure plasma treatment of the substrate improves the Ag nanoparticles (NPs) deposition rate, increases the thin film adhesion as well as the density and compactness of the deposited Ag NPs; leading to the fabrication of pinhole-free thin films (Lisco et al. 2017). The presence of pinholes in the silver thin film surface indicates to the sparse distribution of bacteria along the sensor surface. Therefore, in some parts of the sensor no bacterium is present; resulting in the weaker excitation of the SPs along the metal-dielectric interface. Hence, here, CAP treatment of glass substrate with angle of 60°, treatment time 2 min, applied voltage 10 KV, nozzle-glass distance 1 cm and gas flow rate 2slm was used before deposition.
CAP treatment of the silver thin film deposited on the CAP-treated glass substrate Based on the discussions in Introduction Section, different CAP treatments including 90°(1 min) þ 60°(1 min), 60°(1 min) þ 90°(1 min), and 60°(2 min) were used to decrease the Ag NPs size ; increase the Ag NPs surface area for improvement in binding of silver nanoparticles to bacterial proteins. Based on the WCA measurements, AFM and curve fitting results, among the examined conditions, the most effective CAP treatment condition for surface hydrophilicity and protein adsorption improvement of Ag NPs were obtained to produce the silver-based SPR biosensor. However, the rest of CAP treatment parameters were kept constant (applied voltage 8 KV, treatment time 2 min, applied frequency 6 KHz, nozzle-silver thin film distance 1 cm and gas flow rate 2slm).

Characterization of the glass substrate before and after CAP treatment
Measuring the contact angle of water droplet on the solid surface is used for surface quality control in coating technology (Chen et al. 2018). Before and after cold atmospheric-pressure plasma treatment, the changes in the wettability of glass substrate surface were evaluated using water contact angle measurements at 25°C by placing a 5 μl droplet of distilled water at the glass surface and capturing an image of drop with a camera a few seconds after deposition of drop. To analyze the drop Uncorrected Proof image and calculate the contact angle value, an image analysis program, Image J (Drop analysis-LB-ADSA) was used (Williams et al. 2010).
The surface morphology of the glass substrate was also investigated before and after CAP treatment using AFM by a Nanosurf Mobile S at scanning area 9 μm Â 9 μm.
Characterization of the untreated silver thin film deposited on the untreated glass substrate and CAP-treated silver thin film deposited on the CAP-treated glass Since surface hydrophilicity can be used to control cell behavior by bacterial protein adsorption for sensing application, before and after CAP treatment of silver thin film, the WCA values were measured. For this purpose, at 25°C, 5 μl drops of distilled water was placed at the silver thin film surfaces and captured images a few seconds after deposition of drop.
Before and after CAP treatment, the surface morphology of the silver thin films was recorded in non-contact mode using atomic force microscope (AFM) by a Nanosurf Mobile S at scanning area 9 μm Â 9 μm. The surface roughness was found by calculating the root mean square (RMS) of AFM data and can be evaluated by the following formula (Hosseinpour et al. 2019): where R(l, t) is root mean square, h is average surface height and h(r, t) is height of a point respect to a reference surface at time t whiler is the vector of point place. The average height h for a digitized surface is shown as: Since it has been reported that Ag NPs with smaller the size but larger the surface area have a higher capability of penetrating the bacterial cell membrane than bigger particles (Qing et al. 2018), to measure the size and surface area of silver NPs after CAP treatment, the XY Extract Graph Digitizer v5.1 and Origin Pro v8.6 software were applied for curve fitting. For this purpose, firstly, the 1D AFM profiles were digitized using the XY Extract Graph Digitizer software to find the coordinates of the corresponding grains (Ag NPs) with great accuracy. These obtained dots are considered as digitized data. Secondly, the digitized data of assumed NPs were approximated using the analytical software, Origin Pro, with Gaussian amp function which is defined by the following equation (Verdeyen 1995;Hosseinpour et al. 2019): where h(x) signifies the peak height of grain, the parameter w is standard deviation and x is a position of each point in which peak height is calculated for that point (in μm). Please note that the measured grains were indicated by arrows in 1D AFM profiles. As curve fitting method was performed to analyze the coordinates of the Ag NPs with Gaussian amp function (Equation (3)) using Origin Pro software, this software calculates and reports the values of surface area (nm), height (nm) and FWHM (μm) of the fitted curve on the coordinates of the 1D AFM profile of Ag NP. Additionally, here, the errors obtained by Origin Pro software for the calculated values of surface area, height and FWHM of Ag NP were reported in the corresponding plots in the supplementary material. Field emission scanning electron microscopy (FESEM) of the silver thin film deposited on CAP-treated glass substrate was also carried out (JEOL, JSM 7800F) at an accelerating voltage of 25 KV to show the improvements of thin film adhesion after CAP treatment of glass substrate.

Sensing setup
Design of sensor holder using 3D printing technology Based on the discussion in Section 1, we first designed the 3D structure of the sensor holder using Auto CAD v2015 software. The binder jetting process of 3D printing technology was utilized to fabricate the CAD model with ABS material (Fig.S1 in the supplementary material). The step-by step principle of plotting and designing 3D structure of the chip holder using CAD model (v2015) is shown in Fig.S2.

SPR measurements
A schematic diagram of the experimental setup used for the present study is shown in Figure 2. To prepare bacterial sample E. coli solution was diluted to lower concentrations (10 8 , 10 7 , 10 6 , 10 5 , 10 4 CFU/ml) with distilled water. These bacterial sample solutions with varying concentrations were respectively filled into the flow cells ( Figure 2) using insulin syringe 2CC (cube centimeter). The sensor chips were correctly embedded and fixed in the produced ABS sensor holders (Figs.S1 (b), (c)) by 3D printer and suspended in the spectrophotometer flow cells containing various concentrations of E. coli, respectively. The flow cells were placed in the spectrophotometer instrument with the help of a 3D translational stage (See Fig.S1 (d)). Finally, the SPR measurements were recorded using a spectrophotometer (Perkin Elmer Lambda 25 with an optical resolution of 0.5 nm) based on the intensity-wavelength interrogation methodology (Prabowo et al. 2018) with the help of experimental setup in Figure 2.

Performance parameters of the SPR sensor
The performance of an SPR sensor can be described by several parameters, such as the resonance wavelength, λ res , sensitivity, FWHM, figure of merit (FOM), and resolution. Hence, the sensitivity of the sensors is defined as the following equation (Dormeny et al. 2020): where Dl res is the change in the resonance wavelength and Dn d is the change in the log E. coli concentration (CFU/ml). The average sensitivity of the sensors is also calculated by the slope of a linear fit to the calibration data when the l res (nm) is plotted as a function of log E.coli concentration (cfu/ml) (Dormeny et al. 2020).
The value of the FOM can be calculated by Equation (5) (Meng et al. 2017): where S n is the sensitivity, and FWHM can be determined by calculating the full width at half maximum of the reflectance dip (Dl 0:5 )

Uncorrected Proof
In this work, the wavelength interrogation method was used. Refractive index (RI) resolution (r sensor ) is an important indicator for a wavelength interrogation SPR sensor, which can be affected by signal-to-noise ratio and can be obtained as follows (Ma et al. 2021): where dl is RI detection accuracy, and S n is the RI sensitivity of the SPR sensor, respectively (Rahman et al. 2017a(Rahman et al. , 2017b. dl is mainly affected by the FWHM and reflectance dip of the SPR curve. The standard deviation of the resonance wavelength here is used as the detection accuracy (dl) of the wavelength (Ma et al. 2021).
Stability tests of the untreated and CAP-treated silver-based SPR sensor

Hydrophobic recovery
Hydrophobic recovery is one of the drawbacks of the plasma-treated surfaces. Different authors have stated that the increase in surface hydrophilicity upon plasma treatment is only temporary (e.g. Reche et al. 2016). For our experiment, where the purpose is to enhance the wettability of the silver thin film surface, it is necessary to study the hydrophobic recovery or ageing process. For monitoring the variation in the surface wettability, the WCA values of the CAP-treated silver thin films were obtained over time (24 h : where u S1 is the saturation value of the WCA after CAP treatment, u S2 is the plateau value of the WCA after storage in air and u untreated is the WCA value of the untreated silver thin film.

Reflectivity spectra
The reflectivity spectra of the untreated and CAP-treated silver thin films were recorded in air environment over time (24 h, 48, 72 and 96 h) and the shifts in the resonance wavelength were used as a measure for the optical stability of the produced sensors. For this purpose, prior to the WCA measurements, the reflectivity spectrum for each sample was recorded over time (24 h, 48, 72 and 96 h) using a spectrophotometer (Perkin Elmer Lambda 25 with an optical resolution of 0.5 nm) based on the intensity-wavelength interrogation methodology (Wang et al. 2017;Prabowo et al. 2018).

OES measurement
The emission spectrum of the employed argon plasma jet in the air medium is illustrated in Figure 3(a) for non-perpendicular and in Figure 3(b) for perpendicular incidence angle of jet, respectively. The multiple transition lines were observed in the optical emission spectrum of the plasma jet afterglow operating in ambient air. Accordingly, based on atomic spectra database lines, an intense atomic oxygen line at 777 nm, an intense OH rotational band at 308.5 nm and strong atomic argon emission lines in the range 690-851 nm can be observed. However, some low intensity emissions are also obtained: emission due to the presence of NO radicals at 282.83 nm, emissions attributed to atomic oxygen at 927 nm. Moreover, as shown in Figure 3, lower intensity of OH rotational band at 308.5 nm was observed for non-perpendicular incidence angle of jet compared to the perpendicular incidence angle.

CAP-treated glass substrate characterization
As presented in Fig.S3, the remarkable increment (about 6 nm) in the roughness of the glass surface substrate was observed after CAP treatment with plasma exposure of 60°under the mentioned operating conditions in 'CAP treatment of the glass substrate' Section. Based on the Wenzel model (Nosonovsky & Bhushan 2009) with an increase in roughness (Figure 4(a)), a hydrophilic surface (WCA of glass substrate before CAP treatment ¼ 42 , 90°) becomes more hydrophilic (Figure 4(b): WCA of glass substrate after non-perpendicular CAP treatment ¼ 29:5 ). This CAP treatment of glass substrate leads to the enhancement in the surface hydrophilicity of glass substrate about 12.5°(See Figure 4), resulting in the improvement in the adhesion of thin film to the glass surface with more uniformity of silver nanoparticles; indicating to the production of pinhole-free silver thin film, as shown in FESEM images ( Figure 5).

Wettability
Based on the discussion in 'Introduction' Section, surface hydrophilicity improvement can play important role for protein adsorption; indicating to the bacterial detection (Wigginton et al. 2010) and CAP treatment is a promising technique for  Uncorrected Proof this purpose. According to WCA measurements, as presented in Figure 6(a) and 6(b), a remarkable reduction in the WCA value (about 54.63°) was observed after CAP treatment with 90°(1 min) þ 60°(1 min) . However, insignificant reductions were obtained in the WCA values for the other examined CAP treatments, as shown in Figure 6(c) and 6(d) which can be explained by the results of AFM and curve fitting of Ag NPs using the XY Extract Graph Digitizer and Origin Pro software. As can be seen in Fig.S6 (d), the surface roughness of Ag film treated with 90°(1 min) þ 60°(1 min) exhibits a remarkable increase (about 7 nm) when compared to Ag film treated with 60°(1 min) þ 90°(1 min) . Additionally, compared with the Ag film treated with 60°(1 min) þ 90°(1 min), CAP treatment with 90°(1 min) þ 60°(1 min) can increase remarkably the surface area of Ag NP (Fig.S6(c)) and decrease the FWHM of Ag NP (Fig.S6(b)); which finally leads to the larger enhancement in the hydrophicility of the Ag film treated with 90°(1 min) þ 60°(1 min). The results of OES measurements in Figure 3 indicate that the plasma exposure at 90°shows a slightly higher OH intensity compared to plasma emission at 60°; however, the intensity of OH species is still minor and cannot be the major factor for influencing surface roughness and hydrophicility of Ag film.

Morphological studies with curve fitting for wettability improvement
It has been reported that smaller Ag NPs has a large surface area in contact with the bacterial cellular structures; leading to the surface hydrophilicity or protein adsorption improvement (Wigginton et al. 2010;Polini & Yang 2017).
The Origin Pro software using curve fitting technique has the advantages of calculating the surface area and full width at half maximum (FWHM) for Ag NP size evaluation with the help of 1D AFM profiles shown in Fig.S5(a) (before CAP treatment) and after CAP treatments (Fig.S5(b) and Fig.S5(c)), respectively . Using this technique, as presented in Fig.S6 (a), Fig.S6 (b), and Fig.S6(c), after CAP treatment with 90°(1 min) þ 60°(1 min), the surface area of Ag NPs remarkably increases due to the increment in the height of Ag NPs. However, as it is shown in Fig.S6 (b), the noticeable decrement in the FWHM of Ag NP for the silver film treated with 90°(1 min) þ 60°(1 min) was observed; indicating to the appearance of NPs with smaller size than both untreated silver thin film and CAP-treated silver thin film with 60°(1 min) þ 90°(1 min).  In addition, after CAP treatment of 90°(1 min) þ 60°(1 min), more uniform distribution of Ag NPs with smaller size on the surface was appeared (See Fig. S4).

Comparison between the performances of the untreated and CAP-treated silver-based SPR biosensors
It would be interesting to compare the sensitivities of the CAP-treated pinhole-free silver thin film and the untreated silver thin film deposited respectively on the CAP-treated and untreated glass substrates. Hence, the spectra of the reflected light before immersing the untreated silver thin film SPR biosensor and CAP-treated SPR biosensor inside the E. coli solutions respectively are shown in Figures 7(a) and 8(a). According to Figures 7(a) and 8(a), the strength of SPR signal became weaker at the resonance wavelength for the untreated SPR biosensor due to the presence of pinholes in the silver film surface deposited on the untreated glass substrate, as presented in the FESEM image (See Figure 5(a)). As discussed previously, based on the intensity-wavelength interrogation methodology for SPR measurements the local refractive index slightly changes, giving rise to a change in the SPR resonance wavelength due to the protein adsorption with Ag NPs (Nguyen et al. 2015). Thus, when the concentration of the bacterial sample around the sensor surface increases due to the protein adsorption of bacterial cells with Ag NPs, the SPR resonance wavelength exhibits a red shift, indicating to the detection (Nguyen et al. 2015), as can be seen in Figures 7(b) and 8(b), respectively, for untreated and CAP-treated silver thin films with 90°(1 min) þ 60°(1 min). In addition, as shown in Figures 7(b) and 8(b), the wavelength shifts to higher value with increasing the concentration of E. coli. This indicates to the fact that the biological binding reactions between the Ag NPs and E. coli bacteria increase as the concentration of E. coli increases; which finally leads to the enhancement in the concentration of bacteria on the sensor surface (Nguyen et al. 2015). The variations of resonance wavelength with changes in E. coli concentrations and the corresponding calibration curves are illustrated in Figure 7(c) for untreated silver thin film and in Figure 8(c) for CAP-treated silver thin film, respectively. A good linearity in the calibration curve is illustrated in Figures 7(c) and 8(c); indicating to the applicability of the SPR sensors for detecting the E. coli bacterium. The average sensitivity of the CAP-treated SPR sensor is 4.1 nm=log cfu=ml ð Þ which is calculated using the slope of a linear fit to the calibration curve, as shown in  The performance parameters for the SPR biosensors are listed in Tables 1 and 2, respectively. Results of the performance parameters in Table 1 indicate that as the concentration of E. coli is increased from 10 4 CFU/ml to 10 8 CFU/ml in the untreated silver based SPR sensor, the sensitivity increases from 0.5 nm=log cfu=ml ð Þ to 1.88 nm=log cfu=ml ð Þ , but the FWHM does not change significantly. This leads to the increasing of the FOM. Similarly, the results in Table 2 exhibit that increasing the concentration of E. coli from 10 4 CFU/ml up to 10 8 CFU/ml in the CAP-treated silver-based SPR sensor can increase sensitivity from 1.5 nm=log cfu=ml ð Þto 2.75 nm=log cfu=ml ð Þwhile the FWHM does not change significantly; thus, this leads to the FOM enhancement. Additionally, based on the Equation (5), the CAP-treated silver thin film sensor exhibits a slightly higher average resolution (r CAPÀtreated Ag film ¼ 0:2479 log(cfu=ml ) when compared with the  Uncorrected Proof untreated silver thin film sensor (r Untreated Ag film ¼ 0:2412 log(cfu=ml ). Compared with the CAP-treated silver thin film (pinhole-free SPR sensor: Table 2), the untreated silver thin film (SPR sensor with pinholes) exhibits smaller resonance wavelengths, slightly lower FWHM with also decreasing of the sensitivity; which finally leads to the decreasing of the FOM, as can be seen in Table 1 for the E.coli concentration from10 4 CFU/ml up to 10 8 CFU/ml.
Stability of the untreated and CAP-treated silver-based SPR sensor Figure 9 shows the evolution of the WCA as a function of storage time for the CAP-treated silver thin films, respectively. Figure 10 illustrates the reduction in the WCA values of the CAP-treated silver thin films over the 4 days (96 h) following the treatment (loss in treatment efficiency (%)) obtained using Equation (7) (Deynse et al. 2015). As can be seen in Figures 9 and 10, the CAP-treated silver thin film still shows hydrophilicity (62.56°, 90°) with loss in treatment efficiency of only 35.46% in 96 h from CAP treatment. Additionally, Figure 11(b) indicates that the reflectivity spectra are nearly the same over time (after 24 h, 48, 72 and 96 h of aging) for the CAP-treated silver thin film. This is probably due to the reduction in the surface activation of the CAP-treated silver thin film over the 96 h following the treatment (Figure 12). The chemical composition of the silver thin film surfaces were assessed by Fourier-transform infrared spectroscopy (FTIR) (400-4,000 cm À1 ) from (FTIR spectrometer, BRUKER; ALPHA). Indeed, as can be seen in Figure 12, the intensity of all of the observed absorption bands in the FTIR spectra of the CAP-treated silver thin film was decreased over time (after 24 h, 48, 72 and 96 h of aging); leading to the weaker incorporation of such polar groups as N-H and O-H stretching vibration (alcohol and phenol) at 3,286 cm À1 as well as N ¼ O symmetry stretching (nitro compound) at 1,378 cm À1 on the silver surface. As presented in Figure 11(a), a red shift can be observed in the reflectivity spectrum of the untreated silver thin film after 96 h of aging. So, one can conclude that the CAP-treated silver thin film is more stable in ambient environment than the untreated silver thin film.
Consequently, the results show that the applicability of the CAP-treated silver thin film (pinhole-free SPR biosensor) for identification of the E. coli is better than that of the untreated silver film. Here, two points must be carefully noticed and discussed in the measured experimental data. According to the FESEM images and WCA measurements, the first one is related to the noticeable effect of CAP treatment of the glass substrate on the Ag NPs deposition rate, the thin film adhesion as well as the density and compactness of the deposited Ag NPs; leading to the fabrication of pinhole-free thin films (Lisco et al. 2017); resulting in better bacterial detection based on protein adsorption compared to the untreated silver film containing pinholes (Wigginton et al. 2010). The second point is attributed to the remarkable influence of CAP treatment of this pinhole-free silver thin film on not only Ag NP size but also the distribution of Ag NPs in the CAP-treated glass surface. The AFM and curve  and larger surface area was observed on the silver thin film surface after CAP treatment (pinhole-free biosensor); leading to the sensitivity and FOM enhancement.

CONCLUSIONS
In this study, for the first time, to our knowledge, a pinhole-free silver based SPR biosensor was produced using CAP treatment of silver thin film surface for identification of Escherichia coli in the distilled water. Also, for the first time, a simple and  Uncorrected Proof low-cost sensor holder is fabricated using 3D printer for suspending correctly the produced silver thin film chip in the standard spectrophotometer flow cell containing bacteria solutions. FESEM results show that CAP treatment of the glass substrate with jet angle of 60°(2 min) improves the Ag nanoparticles (NPs) deposition rate and increases the thin film adhesion; leading to the fabrication pinhole-free thin films. The results of AFM, and curve fitting show that CAP treatment of this pinhole-free silver thin film with 90°(1 min) þ 60°(1 min) indicates the appearance of Ag NPs with smaller size and larger surface area compared to untreated silver film deposited on the untreated glass substrate. The pinhole-free silver based SPR biosensor produced with CAP treatment of both glass substrate and silver film shows better sensitivity, FOM and stability compared to the untreated silver-based SPR biosensor when the E. coli concentrations are increased from 10 4 CFU/ml up to 10 8 CFU/ml. This improved performance is attributed to not only the omission of pinholes but also the appearance of more uniform distribution of Ag NPs with smaller size and larger surface area due to the CAP treatment.

DATA AVAILABILITY STATEMENT
All relevant data are included in the paper or its Supplementary Information.