Colloids and Surfaces B: Biointerfaces 117 (2014) 225–232
Contents lists available at ScienceDirect
Colloids and Surfaces B: Biointerfaces journal homepage: www.elsevier.com/locate/colsurfb
Microscale patterned surfaces reduce bacterial fouling-microscopic and theoretical analysis Ravikumar Vasudevan a,∗ , Alan J. Kennedy c , Megan Merritt b,1 , Fiona H. Crocker c , Ronald H. Baney a a
Department of Materials Science and Engineering, University of Florida, Gainesville, FL, USA Badger Technical Services, Vicksburg, MS, USA c U.S. Army Engineer Research and Development Center, Environmental Laboratory, Vicksburg, MS, USA b
a r t i c l e
i n f o
Article history: Received 22 October 2013 Received in revised form 25 January 2014 Accepted 22 February 2014 Available online 4 March 2014 Keywords: Bioinspired Biofilm Topography Antifouling Catheter Nosocomial infection
a b s t r a c t Microscale patterned surfaces have been shown to control the arrangement of bacteria attached to surfaces. This study was conducted to examine the effect of patterned topographies on bacterial fouling using Enterobacter cloacae as the test model. E. cloacae is an opportunistic pathogen involved frequently in nosocomial infections. It is an important model organism to be studied in the context of healthcare associated infections (HAI) and polydimethylsiloxane (PDMS) based urinary catheter fouling. Patterned surfaces, such as SharkletTM , have shown the promise of being a benign surface treatment for prevention of catheter associated urinary tract infections (CAUTI). To the best of our knowledge, inhibition of fouling by E. cloacae has not been demonstrated on microscale patterned PDMS surfaces. In this study, the SharkletTM and smooth PDMS surfaces were used as controls. All pattern surfaces had statistically significantly lower percentage area coverage compared to the smooth PDMS control. A cross type feature (C-1-PDMS), demonstrated the most significant reduction in percent area coverage, 89% (p < 0.01, ˛ = 0.05), compared to the smooth PDMS control and all other patterned test surfaces. Additionally, theoretical calculations show that C-1-PDMS is the only surface predicted to hold the thermodynamically stable Cassie state, which occurs due to trapping air pockets at the liquid–solid interface. Combined the results provide new insights for designing environmentally benign, novel, microscale patterned surfaces for restricting bacterial fouling. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Bacterial biofilm formation forms the basis for other types of biofouling, such as algal spore and barnacle larvae settlement [1]. The human health implications of bacterial pathogenic biofilms are well documented [2–4]. Previous research [5–8] has addressed the application of antifouling surface designs employing low critical surface energy as a potential solution. Reported findings show that biofouling retention strength to surfaces generally follows a decreasing trend with critical surface tension (used as surface
∗ Corresponding author at: Materials Science and Engineering, 549 Gale Lemerand Drive, Room 100 Rhines Hall, University of Florida, Gainesville, FL 32611, USA. Tel.: +1 707 477 5461. E-mail addresses: [email protected], ravivk@ufl.edu (R. Vasudevan). 1 Current address: Laboratory of Translational Cell Biology, Department of Cell Biology, Emory University School of Medicine, Atlanta, GA, USA. http://dx.doi.org/10.1016/j.colsurfb.2014.02.037 0927-7765/© 2014 Elsevier B.V. All rights reserved.
energy) [9]. This relationship was also observed in aquatic biofouling systems; however, lowering surface energy alone is not sufficient to reduce biofouling [10]. An additional example for the use of surface modifications to control biofouling, is the use of patterned surfaces, which were reported to have efficacy for reducing marine biofouling in controlled laboratory settings [11–17]. Encouraged by the earlier studies that reported the possibility of benign patterned surfaces for bacterial fouling control [18–20], the objective of this study was to assess the effect of patterned surfaces on fouling by Enterobacter cloacae, a bacterium responsible for HAIs in the context of CAUTI [21], which to our knowledge has not been studied before. Another objective of this study is to analyze data from the bacterial fouling of patterned surfaces obtained in this study, to understand how patterned surfaces are capable of reducing bacterial fouling suggested previously [22] by taking in to account wetting theory, namely nonwetting state (Cassie–Baxter [23]) to wetting state (Wenzel [24]) transitions observed on patterned surfaces.
226
R. Vasudevan et al. / Colloids and Surfaces B: Biointerfaces 117 (2014) 225–232
Fig. 1. SEM images of the cross patterned surface, C-1-PDMS, (A) top view, (B) perspective view on PDMS. Nomenclature is defined in Table 1.
2. Theory The presence of patterning on a surface leads to the incorporation of air pockets on immersion in to a liquid or in the presence of a liquid droplet on a surface. The stable state of a droplet suspended over air pockets on such patterned surfaces without complete interfacial adhesion of the liquid to the solid surface is defined as the Cassie state [25]. The stable equilibrium state with complete interfacial adhesion following the expulsion of the trapped air is defined as the Wenzel state [24]. The nature of the stability of these trapped air pockets has been subject to theoretical analysis in the literature. A thermodynamic feasibility condition for underwater stability was derived based on a Gibbs free energy analysis of the composite interface [26]. r > rmin =
−1 cos Y
+ f0 1 +
rf
cos Y
(1)
where r is the Wenzel roughness of the topographical surface, rf is the Wenzel roughness of the wetted fraction of the surface, f0 is the wetted fraction of the surface, Y is Young’s equilibrium contact angle for a smooth surface of the same material and rmin is the minimum roughness value for which the submerged topographical surface will sustain the Cassie state. If the thermodynamic criterion is met Eq. (1), then another approach computes the equilibrium condition on the basis of the Laplace pressure across the air water interface, assuming that the interface is pinned at the edge of the microstructural features. Eq. (2) takes into account these wetting parameters to predict the pressure required for a Cassie state to Wenzel state transition [27,28].
PC =
LV L cos(/2 − Y + AC − A
)
standard photolithography process to prepare the PDMS patterned surfaces was used as previously described [30]. The SharkletTM samples (raised pillars) were made using silicon wafer molds provided by Dr. Liwen Jin. The patterns selected for testing comprises 2 hexagonal repeating units (HC-7-PDMS and 11-H-PDMS), 2 cross type repeating (C-1-PDMS and C-5-PDMS) units and 1 sinusoidal ridge type repeating unit (SharkletTM pattern on PDMS) as a control capable of reducing bacterial fouling and the smooth surface (Smooth PDMS) as a control incapable of reducing bacterial fouling without any surface texturing or treatment. The hexagonal repeating units were chosen as baseline simple and highly symmetric patterns (each being inverse of the other) for testing in the similar Wenzel roughness range. The 2 cross type patterns are two size scale versions of the same pattern to compare across both the lower and upper limits of the feature sizes being considered in the study. The intersecting ridges of the cross pattern allows of greater mechanical stability of the features allowing for an exploration of greater range of Wenzel roughness. The SharkletTM pattern is the sinusoidal repeating unit that was shown to be capable of reducing bacterial fouling previously and was included as a control sample. 3.2. Test organism E. cloacae (ATCC 700258) was used in this study, since E. cloacae strains are a normal inhabitant of the human gastrointenstinal tract [21] and can be frequently isolated from soil and water [31,32]. E. cloacae are opportunistic pathogens and have been shown to be a frequent cause of nosocomial infections and infections in immune compromised patients [21]. 3.3. Characterization
(2)
where PC is the difference in pressure across the air–liquid interface, LV is the liquid–vapor interfacial tension, L is the perimeter of the triple phase contact line, cos(/2 − Y + ) is the resolution component of the interfacial tension in the normal direction with being 90◦ for all test samples considered here and (AC − A) is the area over which pressures acts to force the liquid in to capillary. Eqs. (1) and (2) were then used to determine the robustness of the Cassie state for the topographies being considered in this study for additional analysis and interpretation of the results.
The samples tested were characterized using scanning electron microscopy (SEM) (JEOL 6400 and JEOL 5700, JEOL Inc., Peabody, MA, USA) (Figs. 1 and 2), profilometry (Wyko NT 1000 Profiler, Veeco Instruments, Tuscon, AZ, USA), and goniometry (Rame-Hart Instrument Co., Netcong) (Table 2). Characteristic measures such as recessed area fraction and Wenzel roughness factor, r, is a measure of the roughness of patterned surfaces is the ratio of total surface to the unit planar area of measurement [24], were calculated from the carefully measured feature width, spacing and height, given in Table 1.
3. Experimental
3.4. E. cloacae culture and bacterial fouling
3.1. Test materials and surface fabrication
An overnight culture of E. cloacae was grown in 20 ml of Tryptic Soy Broth (TSB) (BD Difco, Sparks, MD) at 25 ◦ C. The various test surfaces were attached firmly onto glass slides by applying a vacuum. The slides were sterilized by soaking them in a 70% ethanol solution for 10 min, rinsed with sterile filtered water, and dried in a sterile petri dish. Then, the slides were aseptically placed into a slide
Polydimethylsiloxane elastomer (PDMS) sheets were chosen on the basis of widespread use in catheters and other medical implant surfaces [29]. This material is well suited to the fabrication procedure employed to produce the microscale patterned surfaces. A
R. Vasudevan et al. / Colloids and Surfaces B: Biointerfaces 117 (2014) 225–232
227
Fig. 2. Top and perspective SEM images of the PDMS test patterned surface for E. cloacae bacterial fouling: (A) and (B) HC-7-PDMS, (C) and (D) 11-H-PDMS, (E) and (F) C-5-PDMS and (G) and (H) SharkletTM patterned surface.
holder inside a 2 l beaker containing 500 ml of TSB. An overnight culture of E. cloacae was used to inoculate the medium to an initial optical density at 600 nm (OD600 ) of 0.0100. The culture was allowed to grow statically for 48 h at 25 ◦ C.
After 48 h the slides were removed using sterile forceps and gently rinsed with sterile water to remove any non-attached cells. The non-treated side of the slides was wiped with ethanol to remove any cells that could interfere with imaging the bacterial
Table 1 Summary of patterned surfaces, including abbreviated designations and measured feature sizes. Pattern type
Designation
Cross pillars (C)
C-1-PDMS C-5-PDMS
Hexagonal pits (HC)
Feature height (m)
Spacing (m)
23 9
21 4
5 2
HC-7-PDMS
7
3
5
Hexagonal pillars (H)
11-H-PDMS
11
3
2
Pillars
SharkletTM
4,8,12,16a
3
2
a
Feature width (m)
The values indicate the different lengths of the features comprising the SharkletTM pattern (Fig. 2G).
228
R. Vasudevan et al. / Colloids and Surfaces B: Biointerfaces 117 (2014) 225–232
fouling. The slides are allowed to dry in a sterile petri dish before staining with SYTO® 9 for 30 min following the manufacturer’s instruction (FilmTracer LIVE/DEAD Biofilm Viability Kit, Invitrogen, Carlsbad, CA). The stain was removed by gently rinsing the slide with sterile water and then the slide was dried in the dark. A sterile water droplet (30 l) was placed on the slide followed by a coverslip. Images of the bacterial fouling were obtained using a Leica TCS NT Confocal Microscope (Wetzlar, Germany) using a 100× oil objective with a FITC filter (480/500 nm, excitation/emission wavelengths). Ten random fields per slide were imaged at a z position of 0.020 m for analysis using ImageJ software (U.S. National Institutes of Health, http://rsbweb.nih.gov/ij/). Images were processed for red-blue-green color, selecting the green split channel. Images were despeckled and analyzed for particles from 5 to infinity pixels, excluding edges and outlines to determine percent area of the surface covered by the bacterium. 3.5. Statistical methods Statistical comparison of the percent area of bacterial fouling coverage was conducted by using MatlabTM (Natick, MA), the normality of the data was tested with the Jarque–Bera test. Levene’s test for homogeneity was performed to detect unequal variances. Kruskal–Wallis non-parametric test was employed at a significance level of ˛ = 0.05, for analysis of variance. Multiple comparisons between groups were performed using Tukey’s post hoc test. 4. Results and discussion 4.1. Characterization Table 1 provides a list of patterned surface designations and feature dimensions as measured by profilometry and as calculated from scanning electron micrographs (using ImageJ). Top and perspective images of patterned surfaces used for the bacteria fouling test are shown in Figs. 1A and 2A, 2C, 2E and 2G and 1B, 2B, 2D, 2F and 2H respectively. Figs. 3–5 provide examples of the confocal microscopy images used for obtaining the bacterial fouling percentage area coverage data. Also, for comparison with the pressures expressed as water column height, predicted to cause transition from Cassie to Wenzel state, the height of the inoculum was observed to be 3.5 × 10−2 m. It is important to note that C-1-PDMS has a greater feature height (19 m) compared to the other patterned surfaces (3 m). Contact angle analysis: contact angle data for all the test samples is given in Table 2. Of the three measures of advancing, sessile and receding contact angles, the sessile drop contact angle is measured and reported most often in the literature. While sessile drop contact angle is certainly an indicator of the hydrophobicity of a surface, defined as surfaces exhibiting water contact angles above 90◦ , it is an intermediate value that is prone to errors and it is more accurate to mention the advancing and receding angles instead [33]. C-1-PDMS, C-5-PDMS, 11-H-PDMS and SharkletTM are pillar type patterns, i.e., the features stand individually whereas, HC-7-PDMS is a pit type pattern which all features are connected forming pit type repeating units. The pillar type patterns have high contact angles with SharkletTM displaying the highest followed by C-1PDMS, 11-H-PDMS, C-5-PDMS, then the pit type HC-7-PDMS and the lowest contact angle was found to be on the smooth PDMS surface. Wenzel roughness and critical pressure data: Wenzel roughness for the test surfaces, given in Table 3, is in the decreasing order of C-1-PDMS, C-5-PDMS, 11-H-PDMS, SharkletTM , HC-7-PDMS and the lowest being the smooth surface with a value of 1. It is interesting to note that as discussed in the theory, the thermodynamic
Cassie state stability condition (rmin ), which allows for the trapping of air between the water and PDMS interfaces is exceeded only in the case of the C-1-PDMS patterned topography. Therefore, the dynamic condition of critical pressure needed to overcome the Cassie state and transition to the Wenzel state may only be applied for C-1-PDMS. 4.2. Bacterial fouling The mean percentage area coverage on the smooth PDMS control surface (Fig. 3A) was found to be 28.3% with a standard deviation of 25.21%. While this indicates high variability, it should be noted that the data for the smooth surface is not normally distributed as verified by the Jarque–Bera test. The distribution also has a skewness of 1.1845 and a kurtosis of 3.2471 showing a long tail, prone to outliers. Also, 50% and 75% of all smooth PDMS percent area coverage data points lie beyond 17.03% and 10.5% respectively with a maximum recorded coverage of 99.6%. Confocal microscopic images of representative planes on the patterned surfaces are provided in Figs. 3–5. As the percentage area coverage data for all the surfaces deviated from normality (Table S2), Kruskal–Wallis non-parametric analysis of variance was performed, followed by a Tukey’s post hoc test for multiple comparison (Table S1). Statistically significant reductions (p < 0.01) were determined on each patterned surface relative to the smooth PDMS control (Figure S1). Furthermore, bacterial growth on the treated surfaces (
Contents lists available at ScienceDirect
Colloids and Surfaces B: Biointerfaces journal homepage: www.elsevier.com/locate/colsurfb
Microscale patterned surfaces reduce bacterial fouling-microscopic and theoretical analysis Ravikumar Vasudevan a,∗ , Alan J. Kennedy c , Megan Merritt b,1 , Fiona H. Crocker c , Ronald H. Baney a a
Department of Materials Science and Engineering, University of Florida, Gainesville, FL, USA Badger Technical Services, Vicksburg, MS, USA c U.S. Army Engineer Research and Development Center, Environmental Laboratory, Vicksburg, MS, USA b
a r t i c l e
i n f o
Article history: Received 22 October 2013 Received in revised form 25 January 2014 Accepted 22 February 2014 Available online 4 March 2014 Keywords: Bioinspired Biofilm Topography Antifouling Catheter Nosocomial infection
a b s t r a c t Microscale patterned surfaces have been shown to control the arrangement of bacteria attached to surfaces. This study was conducted to examine the effect of patterned topographies on bacterial fouling using Enterobacter cloacae as the test model. E. cloacae is an opportunistic pathogen involved frequently in nosocomial infections. It is an important model organism to be studied in the context of healthcare associated infections (HAI) and polydimethylsiloxane (PDMS) based urinary catheter fouling. Patterned surfaces, such as SharkletTM , have shown the promise of being a benign surface treatment for prevention of catheter associated urinary tract infections (CAUTI). To the best of our knowledge, inhibition of fouling by E. cloacae has not been demonstrated on microscale patterned PDMS surfaces. In this study, the SharkletTM and smooth PDMS surfaces were used as controls. All pattern surfaces had statistically significantly lower percentage area coverage compared to the smooth PDMS control. A cross type feature (C-1-PDMS), demonstrated the most significant reduction in percent area coverage, 89% (p < 0.01, ˛ = 0.05), compared to the smooth PDMS control and all other patterned test surfaces. Additionally, theoretical calculations show that C-1-PDMS is the only surface predicted to hold the thermodynamically stable Cassie state, which occurs due to trapping air pockets at the liquid–solid interface. Combined the results provide new insights for designing environmentally benign, novel, microscale patterned surfaces for restricting bacterial fouling. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Bacterial biofilm formation forms the basis for other types of biofouling, such as algal spore and barnacle larvae settlement [1]. The human health implications of bacterial pathogenic biofilms are well documented [2–4]. Previous research [5–8] has addressed the application of antifouling surface designs employing low critical surface energy as a potential solution. Reported findings show that biofouling retention strength to surfaces generally follows a decreasing trend with critical surface tension (used as surface
∗ Corresponding author at: Materials Science and Engineering, 549 Gale Lemerand Drive, Room 100 Rhines Hall, University of Florida, Gainesville, FL 32611, USA. Tel.: +1 707 477 5461. E-mail addresses: [email protected], ravivk@ufl.edu (R. Vasudevan). 1 Current address: Laboratory of Translational Cell Biology, Department of Cell Biology, Emory University School of Medicine, Atlanta, GA, USA. http://dx.doi.org/10.1016/j.colsurfb.2014.02.037 0927-7765/© 2014 Elsevier B.V. All rights reserved.
energy) [9]. This relationship was also observed in aquatic biofouling systems; however, lowering surface energy alone is not sufficient to reduce biofouling [10]. An additional example for the use of surface modifications to control biofouling, is the use of patterned surfaces, which were reported to have efficacy for reducing marine biofouling in controlled laboratory settings [11–17]. Encouraged by the earlier studies that reported the possibility of benign patterned surfaces for bacterial fouling control [18–20], the objective of this study was to assess the effect of patterned surfaces on fouling by Enterobacter cloacae, a bacterium responsible for HAIs in the context of CAUTI [21], which to our knowledge has not been studied before. Another objective of this study is to analyze data from the bacterial fouling of patterned surfaces obtained in this study, to understand how patterned surfaces are capable of reducing bacterial fouling suggested previously [22] by taking in to account wetting theory, namely nonwetting state (Cassie–Baxter [23]) to wetting state (Wenzel [24]) transitions observed on patterned surfaces.
226
R. Vasudevan et al. / Colloids and Surfaces B: Biointerfaces 117 (2014) 225–232
Fig. 1. SEM images of the cross patterned surface, C-1-PDMS, (A) top view, (B) perspective view on PDMS. Nomenclature is defined in Table 1.
2. Theory The presence of patterning on a surface leads to the incorporation of air pockets on immersion in to a liquid or in the presence of a liquid droplet on a surface. The stable state of a droplet suspended over air pockets on such patterned surfaces without complete interfacial adhesion of the liquid to the solid surface is defined as the Cassie state [25]. The stable equilibrium state with complete interfacial adhesion following the expulsion of the trapped air is defined as the Wenzel state [24]. The nature of the stability of these trapped air pockets has been subject to theoretical analysis in the literature. A thermodynamic feasibility condition for underwater stability was derived based on a Gibbs free energy analysis of the composite interface [26]. r > rmin =
−1 cos Y
+ f0 1 +
rf
cos Y
(1)
where r is the Wenzel roughness of the topographical surface, rf is the Wenzel roughness of the wetted fraction of the surface, f0 is the wetted fraction of the surface, Y is Young’s equilibrium contact angle for a smooth surface of the same material and rmin is the minimum roughness value for which the submerged topographical surface will sustain the Cassie state. If the thermodynamic criterion is met Eq. (1), then another approach computes the equilibrium condition on the basis of the Laplace pressure across the air water interface, assuming that the interface is pinned at the edge of the microstructural features. Eq. (2) takes into account these wetting parameters to predict the pressure required for a Cassie state to Wenzel state transition [27,28].
PC =
LV L cos(/2 − Y + AC − A
)
standard photolithography process to prepare the PDMS patterned surfaces was used as previously described [30]. The SharkletTM samples (raised pillars) were made using silicon wafer molds provided by Dr. Liwen Jin. The patterns selected for testing comprises 2 hexagonal repeating units (HC-7-PDMS and 11-H-PDMS), 2 cross type repeating (C-1-PDMS and C-5-PDMS) units and 1 sinusoidal ridge type repeating unit (SharkletTM pattern on PDMS) as a control capable of reducing bacterial fouling and the smooth surface (Smooth PDMS) as a control incapable of reducing bacterial fouling without any surface texturing or treatment. The hexagonal repeating units were chosen as baseline simple and highly symmetric patterns (each being inverse of the other) for testing in the similar Wenzel roughness range. The 2 cross type patterns are two size scale versions of the same pattern to compare across both the lower and upper limits of the feature sizes being considered in the study. The intersecting ridges of the cross pattern allows of greater mechanical stability of the features allowing for an exploration of greater range of Wenzel roughness. The SharkletTM pattern is the sinusoidal repeating unit that was shown to be capable of reducing bacterial fouling previously and was included as a control sample. 3.2. Test organism E. cloacae (ATCC 700258) was used in this study, since E. cloacae strains are a normal inhabitant of the human gastrointenstinal tract [21] and can be frequently isolated from soil and water [31,32]. E. cloacae are opportunistic pathogens and have been shown to be a frequent cause of nosocomial infections and infections in immune compromised patients [21]. 3.3. Characterization
(2)
where PC is the difference in pressure across the air–liquid interface, LV is the liquid–vapor interfacial tension, L is the perimeter of the triple phase contact line, cos(/2 − Y + ) is the resolution component of the interfacial tension in the normal direction with being 90◦ for all test samples considered here and (AC − A) is the area over which pressures acts to force the liquid in to capillary. Eqs. (1) and (2) were then used to determine the robustness of the Cassie state for the topographies being considered in this study for additional analysis and interpretation of the results.
The samples tested were characterized using scanning electron microscopy (SEM) (JEOL 6400 and JEOL 5700, JEOL Inc., Peabody, MA, USA) (Figs. 1 and 2), profilometry (Wyko NT 1000 Profiler, Veeco Instruments, Tuscon, AZ, USA), and goniometry (Rame-Hart Instrument Co., Netcong) (Table 2). Characteristic measures such as recessed area fraction and Wenzel roughness factor, r, is a measure of the roughness of patterned surfaces is the ratio of total surface to the unit planar area of measurement [24], were calculated from the carefully measured feature width, spacing and height, given in Table 1.
3. Experimental
3.4. E. cloacae culture and bacterial fouling
3.1. Test materials and surface fabrication
An overnight culture of E. cloacae was grown in 20 ml of Tryptic Soy Broth (TSB) (BD Difco, Sparks, MD) at 25 ◦ C. The various test surfaces were attached firmly onto glass slides by applying a vacuum. The slides were sterilized by soaking them in a 70% ethanol solution for 10 min, rinsed with sterile filtered water, and dried in a sterile petri dish. Then, the slides were aseptically placed into a slide
Polydimethylsiloxane elastomer (PDMS) sheets were chosen on the basis of widespread use in catheters and other medical implant surfaces [29]. This material is well suited to the fabrication procedure employed to produce the microscale patterned surfaces. A
R. Vasudevan et al. / Colloids and Surfaces B: Biointerfaces 117 (2014) 225–232
227
Fig. 2. Top and perspective SEM images of the PDMS test patterned surface for E. cloacae bacterial fouling: (A) and (B) HC-7-PDMS, (C) and (D) 11-H-PDMS, (E) and (F) C-5-PDMS and (G) and (H) SharkletTM patterned surface.
holder inside a 2 l beaker containing 500 ml of TSB. An overnight culture of E. cloacae was used to inoculate the medium to an initial optical density at 600 nm (OD600 ) of 0.0100. The culture was allowed to grow statically for 48 h at 25 ◦ C.
After 48 h the slides were removed using sterile forceps and gently rinsed with sterile water to remove any non-attached cells. The non-treated side of the slides was wiped with ethanol to remove any cells that could interfere with imaging the bacterial
Table 1 Summary of patterned surfaces, including abbreviated designations and measured feature sizes. Pattern type
Designation
Cross pillars (C)
C-1-PDMS C-5-PDMS
Hexagonal pits (HC)
Feature height (m)
Spacing (m)
23 9
21 4
5 2
HC-7-PDMS
7
3
5
Hexagonal pillars (H)
11-H-PDMS
11
3
2
Pillars
SharkletTM
4,8,12,16a
3
2
a
Feature width (m)
The values indicate the different lengths of the features comprising the SharkletTM pattern (Fig. 2G).
228
R. Vasudevan et al. / Colloids and Surfaces B: Biointerfaces 117 (2014) 225–232
fouling. The slides are allowed to dry in a sterile petri dish before staining with SYTO® 9 for 30 min following the manufacturer’s instruction (FilmTracer LIVE/DEAD Biofilm Viability Kit, Invitrogen, Carlsbad, CA). The stain was removed by gently rinsing the slide with sterile water and then the slide was dried in the dark. A sterile water droplet (30 l) was placed on the slide followed by a coverslip. Images of the bacterial fouling were obtained using a Leica TCS NT Confocal Microscope (Wetzlar, Germany) using a 100× oil objective with a FITC filter (480/500 nm, excitation/emission wavelengths). Ten random fields per slide were imaged at a z position of 0.020 m for analysis using ImageJ software (U.S. National Institutes of Health, http://rsbweb.nih.gov/ij/). Images were processed for red-blue-green color, selecting the green split channel. Images were despeckled and analyzed for particles from 5 to infinity pixels, excluding edges and outlines to determine percent area of the surface covered by the bacterium. 3.5. Statistical methods Statistical comparison of the percent area of bacterial fouling coverage was conducted by using MatlabTM (Natick, MA), the normality of the data was tested with the Jarque–Bera test. Levene’s test for homogeneity was performed to detect unequal variances. Kruskal–Wallis non-parametric test was employed at a significance level of ˛ = 0.05, for analysis of variance. Multiple comparisons between groups were performed using Tukey’s post hoc test. 4. Results and discussion 4.1. Characterization Table 1 provides a list of patterned surface designations and feature dimensions as measured by profilometry and as calculated from scanning electron micrographs (using ImageJ). Top and perspective images of patterned surfaces used for the bacteria fouling test are shown in Figs. 1A and 2A, 2C, 2E and 2G and 1B, 2B, 2D, 2F and 2H respectively. Figs. 3–5 provide examples of the confocal microscopy images used for obtaining the bacterial fouling percentage area coverage data. Also, for comparison with the pressures expressed as water column height, predicted to cause transition from Cassie to Wenzel state, the height of the inoculum was observed to be 3.5 × 10−2 m. It is important to note that C-1-PDMS has a greater feature height (19 m) compared to the other patterned surfaces (3 m). Contact angle analysis: contact angle data for all the test samples is given in Table 2. Of the three measures of advancing, sessile and receding contact angles, the sessile drop contact angle is measured and reported most often in the literature. While sessile drop contact angle is certainly an indicator of the hydrophobicity of a surface, defined as surfaces exhibiting water contact angles above 90◦ , it is an intermediate value that is prone to errors and it is more accurate to mention the advancing and receding angles instead [33]. C-1-PDMS, C-5-PDMS, 11-H-PDMS and SharkletTM are pillar type patterns, i.e., the features stand individually whereas, HC-7-PDMS is a pit type pattern which all features are connected forming pit type repeating units. The pillar type patterns have high contact angles with SharkletTM displaying the highest followed by C-1PDMS, 11-H-PDMS, C-5-PDMS, then the pit type HC-7-PDMS and the lowest contact angle was found to be on the smooth PDMS surface. Wenzel roughness and critical pressure data: Wenzel roughness for the test surfaces, given in Table 3, is in the decreasing order of C-1-PDMS, C-5-PDMS, 11-H-PDMS, SharkletTM , HC-7-PDMS and the lowest being the smooth surface with a value of 1. It is interesting to note that as discussed in the theory, the thermodynamic
Cassie state stability condition (rmin ), which allows for the trapping of air between the water and PDMS interfaces is exceeded only in the case of the C-1-PDMS patterned topography. Therefore, the dynamic condition of critical pressure needed to overcome the Cassie state and transition to the Wenzel state may only be applied for C-1-PDMS. 4.2. Bacterial fouling The mean percentage area coverage on the smooth PDMS control surface (Fig. 3A) was found to be 28.3% with a standard deviation of 25.21%. While this indicates high variability, it should be noted that the data for the smooth surface is not normally distributed as verified by the Jarque–Bera test. The distribution also has a skewness of 1.1845 and a kurtosis of 3.2471 showing a long tail, prone to outliers. Also, 50% and 75% of all smooth PDMS percent area coverage data points lie beyond 17.03% and 10.5% respectively with a maximum recorded coverage of 99.6%. Confocal microscopic images of representative planes on the patterned surfaces are provided in Figs. 3–5. As the percentage area coverage data for all the surfaces deviated from normality (Table S2), Kruskal–Wallis non-parametric analysis of variance was performed, followed by a Tukey’s post hoc test for multiple comparison (Table S1). Statistically significant reductions (p < 0.01) were determined on each patterned surface relative to the smooth PDMS control (Figure S1). Furthermore, bacterial growth on the treated surfaces (
