Introduction

Biofouling, the unwanted accumulation of biomolecules and organisms on a surface, occurs on all man-made materials introduced to the marine environment.1 Biofouling control is important for any such material, especially those used in the construction of ship hulls.2,3,4 Biofilm formation on a ship’s hull increases drag by increasing the surface roughness, which decreases speed and increases fuel consumption.3,4,5 Some attached invertebrates (e.g., barnacles) can also destroy the anticorrosion layer, which requires significant repairs.6,7 Control of biofouling through the application of antibiofouling or biofouling release coatings can save on costs by reducing drag and hull damage.4,6,7

Recent research has focused on coatings that prevent biofouling in an environmentally responsible manner. Antibiofouling coatings work using two main mechanisms: (1) prevention of biofouler settlement (often using a biocidal coating) or (2) reduction of adhesive force between fouling organisms and the surface.6 Biofouling release coatings (BRCs) prevent the adhesion of fouling organisms by creating a surface to which it is difficult to strongly adhere.1,8,9 Novel surfaces with engineered microtextures,3,10,11,12,13 polymer coatings (hydrophilic polymer brushes, elastomeric networks, hydrogels, etc.),1 smart polymers,14,15 and biomimetic materials 16 have been developed to minimize biofilm attachment and allow easy removal of attached biofilm. BRCs are of particular interest to the shipping industry because they use the motion of a vessel through the water to remove fouling organisms.9,17,18 These coatings have a low surface energy and are minimally adhesive, thus preventing the strong attachment of organisms, which helps the vessel move through the water with less drag (improving fuel efficiency).9

Laboratory studies of BRCs have gained interest, given the need to develop environmentally friendly alternatives to biocidal coatings.19 Previous studies have quantified biofilm release using turbulent channel flow apparatuses,20,21,22 spinning water jets,23,24,25,26 a wall jet,27 a handheld water jet,28 and a motorized water jet apparatus.29,30 Each of these devices shares a common mechanism: a flow of water meant to detach fouling organisms from a surface of interest and determine the strength of the organisms’ attachment.20,21,22,23,24,26,27,29,30,31 The calibrated water jet (CWJ, Fig. 1) simulates the motion of a ship through water, and the biofouling release that occurs as a function of ship speed is quantified by determining organism attachment over a surface of interest. Unlike the spinning water jets,23,24,25,26 handheld water jet,28 and the motorized water jet apparatus,29,30 the samples are submerged to mimic the conditions of a ship. Practically, the CWJ is smaller and less complicated compared to the turbulent channel flow apparatus,20,21,22 spinning water jet,23,24,25,26 and wall jet.27 The CWJ was constructed for about $2000 USD, which makes it less costly than some of the other devices.20,21,22,23,24,25,26

Fig. 1
figure 1

CWJ setup: (a) 3D printed block used to set the angle at which the water hits the samples, (b) linear actuator arm, (c) support rails for the water jet, (d) rollers allowing the jet of water to travel over the samples, (e) pipe for water jet flow, (f) nozzle controlling the water jet diameter, and (g) sample stage. Red arrows are used to indicate the labels. Blue arrows in side view 2 show how water travels to the samples

Full size image

The goal of this paper is to demonstrate the use of a calibrated water jet for studying marine biofilm release from surfaces as a function of simulated ship speed. A balance of force and linear momentum is used to relate the pressure exerted by the CWJ on an experimental surface to the theoretical speed of a moving ship. Previous studies used similar techniques to determine the adhesion strength of biofilm to a surface of interest, but this is the first to relate biofouling release to simulated ship speed using a CWJ. This approach is shown to be effective for evaluating experimental surfaces fouled under controlled laboratory conditions and in natural seawater.

Materials and methods

CWJ construction and operation

The CWJ was constructed using common laboratory equipment and calibrated based on patent #US 8,984,958 B1 (Fig. 1). Samples (7.5 × 2.5 cm2 glass slides with and without the coating of interest) were mounted on a slide holder secured to the sample stage (Fig. 1g) and submerged in an aquarium tank (76.5 × 31.5 × 32.1 cm3) filled with sand-filtered seawater pumped directly from Narragansett Bay, Rhode Island (41°29.53’ N, 71°25.12’ W). The flow of seawater to the CWJ was controlled by a pressure regulator (Ayer Sales, Inc., product number 7705). The diameter of the jet (7.5 mm for Scenario 1 and 6.0 mm for Scenario 2) was controlled using brass nozzles (Fig. 1f) that were positioned 2.5 cm (1 inch) above the samples. The angle at which the jet of water hits the samples was controlled by an interchangeable 3D printed block (Fig. 1a). For the experiments described here, the jet was set 90° to the samples due to the limited number of Ulva spores available. In coastal Rhode Island and some marine temperate areas, successful harvesting of Ulva quadriflagellated spores is limited to summer months and influenced by the lunar cycle.32 A linear actuator (Firgelli® Automations, part number FA-04-12-3, Fig. 1b) powered by a 12 V battery moved the jet of water for a single pass at a constant 4 cm/s over the surface of the samples. This ensured that the samples were exposed to the stream for the same amount of time.

A calibration curve relating the line pressure and the pressure experienced at the sample surface was generated to account for the difference in pressure between the nozzle output and the surface of the submerged samples. A pressure transducer (Omega Dyne PX309-030GV) was screwed into a small port on the bottom of a tank that was filled with seawater. The nozzle of the CWJ was placed 2.5 cm above the pressure transducer. The pressure in the line gauge was varied from 0 to 200 kPa, and the resulting pressure at the transducer was recorded to create a calibration curve.

Scenario 1: fouling with Ulva under laboratory conditions

All glass microscope slides were cleaned as in Finlay et al.29 Sample slides (n = 24) were coated twice with the marine polish HullKote (Team McLube®), according to the manufacturer’s instructions. HullKote is made of a polytetrafluoroethylene (PTFE) suspension with a citrus-based polish, silane polymer, and a biodegradable detergent.33 The coating also contains aluminum-based particles as cleaning abrasives.33 Uncoated glass slides (n = 24) were similarly cleaned for comparison with the HullKote samples.

Mature, brown tipped Ulva cf. linza blades were collected from Fort Wetherill, Jamestown, RI (41° 28.75′ N, 71° 21.71′ W) and Beavertail Point, Jamestown, RI (41° 27.11’ N, 71° 23.77’ W), on June 13, 2014, and July 9, 2014, respectively, and returned to the laboratory on ice. Ulva blades were rinsed with sterile, 0.45 µm filtered seawater and stored on absorbent paper at 4 °C overnight.34 The blades were sorted into individual tubes with 4 mL of sterile filtered seawater to induce spore release.44 Spore suspensions containing quadriflagellated asexual spores (Fig. S1) were filtered through a 10 µm plankton net and pooled. The concentration of zoospores was determined using a Fuchs–Rosenthal hemacytometer. The June 13th and July 9th samplings resulted in suspensions with 1.8 × 105 and 5.1 × 105 spores/mL, respectively. Each glass slide (HullKote and uncoated) was distributed into quadriPERM® culture dishes, covered with 10 mL of the spore suspension, and incubated at 20 °C for 17 h. A single pass with the CWJ was used to generate biofouling release for each sample. The pressure at the sample surface ranged from 0.69 (hydrostatic pressure only) to 69.90 kPa for antifouling coated slides and 74.63 kPa for uncoated (four slides tested per given pressure). Spores remaining on a given slide were enumerated and averaged from 20 fields of view (each 0.066 mm2) along the path of the CWJ and avoiding edges using a Nikon Eclipse 80i in phase contrast microscopy. Each slide was considered a replicate, and the average number of spores per slide was adjusted to the original suspension concentration to obtain the proportion of remaining spores.

Scenario 2: fouling in natural seawater

Cleaned glass slides (n = 4) were suspended vertically in a 24.3 m3 tank with flow-through natural seawater (Fig. S2) from October 6, 2021, to November 4, 2021. Seawater (Table S1) flows through the tank at a rate of 200 ± 20 L/min, and the tank undergoes ~ 12 turnovers per day. Three random fields of view (each 11.1 mm2) within the path of the CWJ were photographed for each slide using a stereomicroscope (Olympus SZX-16) in oblique light with an AVT-Stingray camera. The area of biofilm coverage was measured using ImageJ software (https://imageJ.nih.gov) before and after a single pass with the CWJ at a constant pressure of 11.6 kPa. A single pressure was chosen for this experiment because we were more interested in determining the effectiveness of this system for evaluating natural biofilm release rather than examining biofilm release at several pressures.

Imaging data analysis

In Scenario 1, normality and outlier testing was performed on arcsine square-root-transformed data using Shapiro–Wilk goodness of fit and quantile range outlier functions in JMP Pro 16.2.0, respectively (Supplemental Methods C). Because data in Scenario 1 were not normally distributed, the nonparametric Kruskal–Wallis analysis followed by the Wilcoxon comparison was used to examine slide to slide variations and differences among pressure conditions. Data on percent coverage in Scenario 2 were arcsine square-root-transformed and were found to be normally distributed, leading to a nested ANOVA test. There were no outliers. The results of the statistical analyses for each scenario can be found in Supplemental Information.

Expression relating measured plate/hull surface pressure to ship velocity

The goal of the following analysis was to develop an equation that can be used to relate the pressure acting on a sample plate to the velocity (Vs) that would produce an equivalent pressure on the hull of a ship. The term Vs is a velocity vector, having a magnitude of ship speed accompanied by the direction of the ship in the positive x-coordinate. Consider a stationary, horizontal, submerged, sample plate on which there is biofilm formation. The plate is submerged in a fluid, assumed to be seawater. To assess biofouling release from the sample plate, seawater flows at a velocity (Vs) in the form of a jet to strike the plate at an angle θ = 90  (Fig. 2a). The practical orientation for the mathematics requires rotating this experimental setup by 90, such that the plate, assumed to be the external surface of the ship’s hull, is oriented normal to the flow of the seawater (Fig. 2b). The assumptions are that the seawater flow is steady with a smooth, submerged plate/hull surface at an angle θ = 90  with respect to the horizontal, and a depth below the fluid surface H. The normal or 90° orientation is the same as that used for the experimental sample plates and flow direction. It is important to note that while the experiments were performed with the force of the fluid striking the sample plates at a directional angle of 90°, there is an additional force acting on the sample surfaces that aids in removing biofilm which was not accounted for in the following calculation. This is the force Ff due to friction or drag acting along or parallel to the hull/plate surface. Including the contribution of Ff, in the determination of the estimated ship velocity of equation (1), however, would not accurately depict the experimental setup. This exclusion though provides an estimation of required ship speed that is more conservative in removing biofilm formation, than if Ff had been included. This is instantly apparent when examining the expression for Ff, as derived in Supplemental Information equation (11), since it is comprised of the cosine of the angle θ in Fig. 2a. The cosine of 90° for the fluid striking the hull/sample surface is zero, making the Ff component nonexistent. However, assume any other angle 0 ≤ θ < 90 degrees, and the Ff component becomes nonzero and adds an additional force component aiding in the biofilm removal from the surface. Also, there is assumed a uniform atmospheric pressure above the seawater surface. Excluded from the calculation are the negative pressure gradients downstream of the submerged plate/hull surface, boundary layer effects such as circulation (i.e., vortices) contributing to drag, effects of transitional flow regimes in the boundary layer, force due to friction or drag, and wave and wake interactions.

Fig. 2
figure 2

(a) Experimental setup of a jet of seawater with a velocity Vs striking a sample plate at an angle θ = 90  to the plate, (b) rotated experimental setup of a jet of seawater with a velocity Vs striking a sample plate at an angle θ = 90  to the plate, for practical calculations

Full size image

The velocity Vs from the jet onto the stationary sample plate of the experimental setup is first transformed as the velocity Vs of the seawater striking the hull of the ship. This arrangement is similar to that of a typical water tunnel experiment where the object, here the ship, is stationary and the seawater flows normal to it at a velocity Vs. The arrangement of Fig. 2b may be transformed once again since it is the same mathematically as the ship having a velocity Vs and moving through a static fluid in the opposite direction.

The details of the force and linear momentum balances required to establish the relationship between the water pressure onto the sample surface and ship speed are presented in Supplemental Information, along with assumptions. The final equation is:

$${V}_{S}=\sqrt{\frac{{P}_{m}+{\gamma }_{{\text{seawater}}}*H}{\rho }}$$
(1)

where Vs is the ship velocity (m/s); Pm is the pressure measured at the sample surface (kPa); γseawater is the specific weight of seawater (N/m3); H is the distance below the fluid surface at which the centroid of the vertically oriented plate/hull surface is located (m); and ρ is seawater density (kg/m3).

Results and discussion

Scenario 1 results

Pressure at the sample surface [Pm in equation (1)] is used to simulate the conditions on the ship hull during operation. Simulated ship speeds were varied 2.0–9.0 m/s using the CWJ. Figure 3a shows that Ulva spore detachment increased with increasing pressure Pm, thus with increasing simulated ship speed, in comparison with no speed or simple hydrostatic pressure (2.0 m/s or 0.69 kPa) for HullKote samples. The density of spores remaining on the surface significantly decreased with increasing water jet pressure (p < 0.0001; R2 = 0.75). This Ulva release tended to be higher for the HullKote (56–82% release compared to hydrostatic pressure) than for the uncoated glass slides (40–59% release compared to hydrostatic pressure). These values are within the published range for water jets operated with pressures from 20 to 110 kPa.25,26,31,35 Less biofouling was released from HullKote coated samples compared to some coatings (e.g., 64% less than polydimethylsiloxane with copper–carbon nanoparticles at 20 kPa35), but HullKote also performed better than others (e.g., 23% greater biofouling release than silastic coatings at 70 kPa23). The concentration of remaining spores on uncoated glass slides did not change significantly with increased water jet pressure (p > 0.19; R2 = 0.09). This is partially a result of large variations in spore concentrations from slide to slide following one pass of the water jet at each pressure (Fig. 3b). Studies examining biofouling release using a similar water jet apparatus either do not submerge their samples or do not report the depth at which they are submerged so calculation of simulated ship speed is not possible.25,26,31,35

Fig. 3
figure 3

Fraction of Ulva spores remaining on (a) glass slides coated with HullKote and (b) uncoated glass slides (n = 4 slides averaging 20 fields of view per pressure at the surface Pm) after one passage of the water jet. Spore counts per slide were adjusted for the initial concentration of spores. In both (a) and (b), the nonparametric Kruskal–Wallis test followed by the Wilcoxon comparison assigned the significant differences (lower case letters) in the data. No outliers were detected in (a) or in (b)

Full size image

Scenario 2 results

Natural biofilm formation on slides exposed to flow-through seawater for four weeks consisted primarily of algal material and detritus.36 There was a proliferation of pennate diatoms (single, colonies, on stalks, and tube dwelling) and some seaweed germlings (Fig. 4b). The CWJ removed 79% of the mature biofilm (Fig. 4) at a simulated ship speed of 3.75 m/s (11.6 kPa at 28 cm).

Fig. 4
figure 4

Naturally developed biofilm from glass slides before and after CWJ testing. (a) Biofilm coverage (n = 15); example of a slide surface before (b) and after (c) one pass of the CWJ. Error bars represent one standard error about the mean of all fields of view. Lower case letters within panel (a) indicate groups of statistically different conditions (p < 0.0001)

Full size image

CWJ performance

The speeds simulated by the CWJ [2.0–9.0 m/s, calculated using equation (1)] were on the lower end of the range of speeds capable for personal boats (4–25 m/s)37,38 and lower middle range for larger container ships and oil tankers (5–13 m/s).39,40,41 This range of speeds was calculated from the sample depth during the experiments (28 cm). Typically, large ships and tankers have hulls submerged to a depth of 3–10 m.42 If equation (1) is modified for this depth, the pressures applied experimentally would be appropriate for simulating vessel speeds of 5–13 m/s. This is accurate provided the internal Froude number for the flow is greater than 1, signifying that subsurface wave effects are unimportant.43

The scenarios described here were designed to examine the performance of the CWJ under a variety of experimental conditions. In Scenario 1, HullKote coating has greater biofouling release with increasing CWJ pressure (Fig. 3) than uncoated slides. The increase in biofilm release for HullKote with simulated ship speed is as expected based on the force and linear momentum balance information presented above. The relationship between ship speed and biofouling release complements the work of previous studies which have measured the effect of shear stress on biofilm removal.20,21,22,23,24,27,29,30

In Scenario 1, Ulva spores were applied as a model biofouler under laboratory conditions. The applicability of the CWJ for quantifying biofouling release from naturally fouled surfaces was demonstrated in Scenario 2. The results from the two testing scenarios indicate the applicability of the water jet for quantifying biofouling release from BRCs fouled under different conditions. By using equation (1) and a similar CWJ apparatus, researchers can examine the performance of BRCs under ship speeds similar to what would be experienced during operation. The CWJ demonstrated applicability for different fouling scenarios, indicating that it could be utilized for different experimental protocols, irrespective of seasons or test site geography, thus contributing to a better standardized reporting for coating comparability.44

Scaling analysis

The experiments performed in this study were conducted using glass slides which have a relatively small surface area compared to ships. Dimensional analysis was employed to provide a scaling law where the measured normal pressure from small-scale experiments to remove biofilm is used to predict the required velocity of a large-scale ship to perform similar biofilm release from the hull (Fig. 5).45 Using the parameters of the experiments, fluid properties and ship velocity, the Buckingham Pi theorem was employed to produce dimensionless groups that allow for correlation between the measured pressure Pm at the plate/hull surface and the ship speed Vs.45 For γ = specific weight of seawater, Lp the characteristic sample length, and Reynolds number Re, the most useful Pi groups are,

Fig. 5
figure 5

Scaling relationship between Vs and Pm. Black dots indicate the values used for the scaling analysis. The solid blue line is the linear regression of the points. The red dashed line is the power law regression for the points. The equations written in red (top left) or blue (bottom right) correspond to the power and linear regressions, respectively

Full size image
$$\frac{{P_{m} }}{{\gamma \cdot L_{p} }},\frac{{V_{s} ^{2} }}{{gL_{p} }}{\mkern 1mu} {\text{and}}{\mkern 1mu} \frac{{\rho V_{s} L_{p} }}{\mu } = Re$$

From this analysis, curve fits are produced such that for a linear curve fit (solid blue line), the measured pressure Pm was found to be proportional to the ship velocity Vs to the power of 2, and for a power law curve fit (red dashed line), to the power of 2.8 (Fig. 5).

Using Pi groups, the linear model describes the relationship between measured pressure and estimated ship velocity of equation (1), such that the measured pressure Pm correlates with the square of the ship velocity, Vs. The power law model raises the Vs2 term to the 1.53 power, making the measured pressure correlated with Vs2.8. This is larger than the estimated ship speed correlation of Vs2 and has a lower coefficient of determination, R2, than the linear fit. The scaling analysis is valid for Reynolds numbers (Re) in the range 1E4 ≤ Re ≤ 1E6, as shown on the right-hand vertical axis of Fig. 5, and typical of a range of Reynolds numbers for fluid flow related to large ships. Overall, this analysis indicates that the small-scale experiments correlate well with the larger-scale ship velocity.

Future work

Equation (1) allows one to relate the results of CWJ experiments to a theoretical ship speed. It could be applied to estimate how much pressure and biofouling release occur during ship operations and would inform the frequency of hull cleaning. While this relationship is useful, there are some limitations for this equation. Equation (1) assumes that the plate/hull surface is perpendicular to the flow of the water; thus, the equation only models the pressure at the bow. At any other angle, water flows past the plate/hull surface and shear forces would be dominant. This would be similar to the shear stresses modeled in studies utilizing a wall jet27,29 or turbulent flow chamber.20,21,22 However, the angle at which the water strikes the surface in the CWJ can be adjusted by modifying the 3D printed block that holds the tubing in place (Fig. 1a), so future work should examine how to adapt this equation to other experimental designs.

Conclusions

In this study, the theory and application of a water jet calibrated for water pressure were examined for potential application in the study of biofouling release from surfaces exposed to the marine environment. A force and linear momentum balance was used to develop an equation relating pressure at the sample surface to simulated ship speed. The CWJ technique also showed applicability for studying biofouling release using surfaces fouled with a model marine biofouler as well as through submersion in natural seawater. This demonstrates a wide applicability for the CWJ in different testing scenarios, making performance analysis of BRCs easier. Utilizing the CWJ and equation (1) would be an evaluation step toward commercial application of different coating types by determining fouling release under simulated operating conditions. Overall, the results presented here can assist in testing novel antifouling coatings for application in industry.