Comparison of laboratory and field testing performance evaluations of siloxane-polyurethane fouling-release marine coatings

Abstract A series of eight novel siloxane-polyurethane fouling-release (FR) coatings were assessed for their FR performance in both the laboratory and in the field. Laboratory analysis included adhesion assessments of bacteria, microalgae, macroalgal spores, adult barnacles and pseudobarnacles using high-throughput screening techniques, while field evaluations were conducted in accordance with standardized testing methods at three different ocean testing sites over the course of six-months exposure. The data collected were subjected to statistical analysis in order to identify potential correlations. In general, there was good agreement between the laboratory screening assays and the field assessments, with both regimes clearly distinguishing the siloxane-polyurethane compositions comprising monofunctional poly(dimethyl siloxane) (PDMS) (m-PDMS) as possessing superior, broad-spectrum FR properties compared to those prepared with difunctional PDMS (d-PDMS). Of the seven laboratory screening techniques, the Cellulophaga lytica biofilm retraction and reattached barnacle (Amphibalanus amphitrite) adhesion assays were shown to be the most predictive of broad-spectrum field performance.


Introduction
The development of new coatings and surfaces to control marine growth requires the implementation of tests that challenge the surfaces in a manner that enables a scientific understanding of the mechanisms by which they function and in a manner that is representative of the environment for which they are being designed. The former is best accomplished through rigorous laboratory testing where both the environments and organisms can be controlled. The interactions between material surfaces and the marine environment, however, are complex due to differences in chemistry, physical conditions and biology (Swain & Schultz 1996;Swain 1997). For this reason, the ultimate test is to expose the materials to static and dynamic immersion at field test sites located at different geographic locations.
Research at North Dakota State University (NDSU) has taken a combinatorial approach to the development of new fouling control coatings Chisholm et al. 2008;Webster et al. 2009;Webster & Meier 2010). This approach includes rapid biological assays that involve exposing the coatings to marine bacteria, algae (diatoms and macroalgae (Ulva linza)), and barnacles (Amphibalanus amphitrite). The hypothesis was that a suite of laboratory assays can be used as a predictor of coating performance in the field, allowing for downselection of coatings for field testing from a large number of experimental candidates. This paper analyzes the laboratory and field data that were obtained for a series of hybrid siloxane-polyurethane fouling-release (FR) coating systems that enable the efficient release of biofouling while providing improved toughness and durability when compared to traditional, elastomeric technologies Majumdar et al. 2007;Pieper et al. 2007;Sommer et al. 2010;Bodkhe et al. 2012). Coatings based on this hybrid strategy have predictability to the performance observed in the field. A number of studies have been published in recent years that report on the degree of correlation or the relationship between performance data garnered from biological laboratory assays and static ocean immersion testing. In this regard, the authors have previously shown that rapid laboratory assays based on marine bacteria, diatoms and barnacles agreed very well with static ocean immersion performance assessments of AF polysiloxane coatings containing tethered biocides (Webster et al. 2009) and polysiloxane FR coatings (Stafslien et al. 2007a;Rittschof et al. 2008;Stafslien et al. 2012). Phang et al. (2009) reported that barnacle cyprid settlement behavior in the laboratory was an effective indicator of short-term macrofouling colonization on functionalized-glass surfaces in the field. Likewise, Martinelli et al. (2012) showed that cyprid settlement and macroalgal sporeling removal assays functioned as good general predictors of static field immersion testing (18 weeks of exposure) for a series of poly(dimethyl siloxane) (PDMS)-based amphiphilic copolymer blends, while Zhang et al. (2013) established strong correlations between diatom and macroalgal-based laboratory testing and static and dynamic assessments of commercial FR coating systems in the field for immersion up to 236 days. Efimenko et al. (2009) showed that macroalgal sporeling removal from hierarchically wrinkled surfaces agreed well with diatom adhesion assessments collected in the field, but not with macrofouling removal observed after immersion for four months. A study by Bressy et al. (2010) resulted in similarly mixed results for a set of ammonium salt-based paints, which exhibited good AF activity in early-stage field assessments (exposure for one month) and against barnacle cyprids in the laboratory, but not towards bacteria and diatoms.
The motivation for the present study was twofold: (1) to assess the FR performance of a set of novel siloxane-polyurethane coatings in both the laboratory and in the field; and (2) to examine the degree of correlation between these two datasets at both the individual assay/assessment level and in the context of overall FR performance. To accomplish these objectives, the siloxane-polyurethane coatings were prepared in multi-well plates and on 10 cm × 20 cm panels. Laboratory screening assessments of bacteria, microalgae, macroalgal spores, and adult barnacle and pseudobarnacle adhesion were conducted, as were evaluations of slime and soft and hard fouling adhesion, conducted at three static ocean immersion testing sites after exposure for one to six months. This paper provides a comprehensive analysis of these FR performance assessments and describes the statistical relationships amongst them. The strengths and weaknesses of each laboratory screening assay in the context of predicting broad-spectrum field performance are also discussed. been shown to perform comparably to commercial FR coating systems in laboratory assays while possessing improved toughness (Sommer et al. 2010). However, the existing database of FR properties collected for the siloxane-polyurethane coatings comprises laboratory testing only and assessments of their performance in a real-world marine environment are necessary to fully vet their potential as effective FR coatings.
Laboratory testing offers the ability to evaluate coatings against specific organisms under controlled conditions on demand. To this end, a variety of assays have been developed and applied to screen antifouling (AF) and FR coating performance towards both microfouling and macro-fouling organisms (Briand 2009). These assays typically utilize small-sized samples (ie microscope slide or coverslip geometries), accommodate the evaluation of large numbers of coatings in parallel and can be completed in several days or weeks under precisely controlled conditions. The majority of the assays employ a single fouling organism to enable species specific information regarding AF/FR performance to be obtained. In most cases, laboratory-based techniques are unable to provide information regarding multi-organism dynamics/interactions or longterm performance data (ie activity, durability) that may be gleaned from static ocean immersion trials.
Static ocean immersion testing typically involves the deployment of coated panels (eg 25 by 30 cm) from a stationary platform or raft at a relatively shallow depth of water (eg 1 m) for several weeks to months. During this time, the coated panels are periodically removed and assessed for the accumulation and adhesion of biofouling to the coating surfaces using standardized testing methods (Lim et al. 2014). These types of performance evaluation provide valuable information with regard to a coating's broad-spectrum AF/FR activity and long-term durability in an aggressive, real-world environment. Furthermore, they do not require as much material and long duration of operational deployment required to evaluate coating performance through large patch testing on a ships' hull. However, static ocean immersion testing may be susceptible to the seasonal fouling activity at the specific test site and cannot control for specific environmental conditions.
The methodology(s) employed to assess the efficacy of new AF/FR coating technologies is largely dictated by constraints unique to each coating developer, such as the available resources (ie labor, finance and materials) and project timelines. Laboratory assays and field testing are often partnered together to initially downselect candidate technologies and identify promising leads, respectively (Sanchez & Yebra 2009). To increase the odds of success of this strategy, laboratory-based screening methods should ideally exhibit some level of

Preparation of siloxane-polyurethane coatings
The preparation of the siloxane-polyurethane coatings has been described in previous publications (Ekin & Webster 2006;Pieper et al. 2007;Sommer et al. 2010). In addition, a number of standards were prepared including Intersleek® 700 (IS700), Intersleek® 900 (IS900, AkzoNobel International Paint, Gateshead, UK), Silastic T2 (T2, Dow Corning, Midland, MI, USA), and a polyurethane not containing siloxane (PU). It is important to note that the latest innovation in the Intersleek® product line, Intersleek® 1100 SR, was not commercially available when this study was carried out, but has been utilized in subsequent fouling studies by the authors to ensure that the FR performance of experimental coatings are adequately measured against state-of-the-art, commercial technologies in both the laboratory and the field. Detailed descriptions of the preparation of the siloxane-polyurethane coatings and standard coatings are provided in the Supplemental material.

Characterization of siloxane-polyurethane coatings
Surface analysis of the coatings was done using contact angle measurements of water and methylene iodide using a Symyx/First Ten Ångstroms™ Coatings Surface Energy System (Symyx, Santa Clara, CA, USA). Surface energy (SE) was calculated using the Owens-Wendt method (Owens & Wendt 1969). Pseudobarnacle (PB) adhesion is the removal force required to remove epoxy-glued studs from the surface of the coatings. It was measured using a Symyx® Automated Pull Off Adhesion station ). Further details are in the Supplemental material.

Laboratory assessment of FR performance
Bacterial (Cellulophaga lytica) biofilm adhesion was assessed by quantifying both the degree of retraction (percentage surface coverage) and the amount of biofilm removed (percentage removal) after exposure to hydrodynamic shearing generated with an automated water-jetting apparatus (Stafslien et al. 2006;Stafslien et al. 2007a;Stafslien et al. 2007b;Ribeiro et al. 2008;Stafslien et al. 2012;Bodkhe et al. 2015). Likewise, the adhesion of microalgal cells (the diatom Navicula incerta) and macroalgal sporelings (young plants of Ulva linza) (expressed as percentage removal) were quantified using analogous water-jet based methodologies (Callow et al. 1997;Cassé et al. 2007). A reattachment assay was used to quantify the adhesion strength in shear (MPa) of adult barnacles (Amphibalanus amphitrite) 14 days after underwater reattachment (Rittschof et al. 2008;. Details are provided in the Supplemental material.

Field testing assessment of FR performance
The coatings were sent to three field sites with very different environmental profiles for evaluation: Morro Bay, California, the Indian River Lagoon, Florida, and Singapore. The performance was evaluated over a period of six months with respect to the development of fouling community structure and fouling adhesion as quantified via water-jet (pressure to remove fouling) and hand-held force gauge (adhesion strength in shear (MPa)) based methods (Swain & Schultz 1996). Coatings at Indian River Lagoon and Singapore were immersed under caged conditions to prevent predation and grazing which has been observed to alter the results on FR coatings (Swain et al. 1998). Coatings at Morro Bay were not caged. Specific details of the methods used at each test site are provided in the Supplemental material.

Statistical analysis
For the laboratory characterization of FR properties, statistical analysis was performed using JMP 7.0.2, SAS Institute Inc (Cary, NC, USA). One-way ANOVAs were used to evaluate the differences in water contact angles (WCA), surface energies (SE) and adhesion of pseudobarnacles and the marine fouling organisms to the siloxane-polyurethane and standard control coatings. The p-values were reported and a Tukey-Kramer HSD post hoc test was used to compare individual coatings within each dataset (α = 0.05).
The FR performance rankings (1-10; 1 = best, 10 = worst) for each coating in the laboratory assays and field testing assessments were assigned as a function of their 'mean rank score' values calculated from a non-parametric One-way ANOVA (Kruskal-Wallis (rank sums) test; α = 0.05) using JMP 7.0.2, SAS Institute Inc. The Spearman's rank correlation coefficients (r s ) for laboratory screening assays, field testing assessments and laboratory screening assays vs field testing assessments were determined using a 95% confidence interval. Table 1 summarizes the compositions of the cured coatings, showing the variations in the type of polyol and PDMS as well as PDMS content. The molecular weight of both types of PDMS was the same.

Properties of the coatings
The WCA and SE of the cured coatings are shown in Figure 1A. The siloxane-polyurethane coatings showed high WCA, where the mean WCAs were in the range of 105-111° for all coatings. The SE of the cured coatings were also low (14-23 mN m −1 ), as expected for these types standards is shown in Figure 1B. The PB adhesion force for the siloxane-polyurethane coatings was similar for all of the compositions (7-10 N), except for PCL-D20 and ACR-D20, which exhibited a significantly higher removal force of 19 N and 17 N, respectively (ANOVA p < 0.0001). The six siloxane-polyurethane compositions that were determined to be statistically equivalent to one another (shared letter 'C' in Figure 1B) also exhibited comparable PB removal forces to the commercial FR standards (9 N) and were significantly lower than the PCL-PU control (90 N) (ANOVA p < 0.0001).

Bacterial biofilm retraction and removal
The results of the C. lytica biofilm retraction analysis on the siloxane-polyurethane coating surfaces are shown in Figure 2A as a function of percentage surface coverage. Representative images of the biofilm growth and retraction obtained on each coating surface have also been provided in Figure S1A. While most of the siloxane-polyurethane coatings showed retention of the C. lytica biofilm, the biofilm was not uniformly distributed (ie exhibited retraction) on the M-series and IS 900 coating surfaces, indicating a reduced degree of adhesion (Ribeiro et al. 2008). Specifically, the coatings prepared with m-PDMS induced a substantial degree of biofilm retraction (30-60%) when prepared with either the PCL or acrylic (ACR) polyol, as indicated by a reduced surface coverage. However, no biofilm retraction (ie 100% surface coverage) was evident for the compositions prepared of systems, showing that the PDMS had stratified to the coating surface. Interestingly, the SE for the coatings prepared with difunctional-terminated PDMS (d-PDMS) (approximately 15 mN m −1 ) was lower than those prepared with monofunctional-terminated (m-PDMS) (~22 mN m −1 ). The polyol did not affect the WCA or SE of the coatings. The WCA of the polycaprolactone (PCL) polyurethane (PCL-PU) control (82°) (without PDMS) was considerably lower than the coatings prepared with either type of PDMS. The SE of the control (41 mN m −1 ) was also much higher than for the siloxane-polyurethane coatings. Since there were no clear trends in the WCA values as a function of coating composition, these data were not used for further analysis. The difference in SE observed as a function of PDMS type arose mainly from differences in the MI contact angle values, which are a measure of the non-polar component of the surface energy. However, the molecular origin of these variations is not clear and much more detailed surface composition analysis is needed and is part of ongoing research. The pseudobarnacle (PB) adhesion for these coatings and commercial FR  removal was observed on the coating comprising 20 wt% d-PDMS (ANOVA p < 0.0001). The removal of H. pacifica biofilm from the coating surfaces after water jetting at 111 kPa for 5 s is shown in Figure 2C (representative images in Figure S1C). It is important to note that higher water jetting pressures were utilized for the H. pacifica biofilm adhesion assessments, when compared to C. lytica, as laboratory cultivated biofilms of this bacterium tend to adhere more tenaciously to siloxane-based coating technologies. The PCL-PU control and PCL-M20 exhibited the highest amount of biofilm removal among the coatings, achieving 74% and 67% removal, respectively, while the IS 700 and IS 900 standards showed essentially no removal of the attached biofilms (< 5%). Coatings PCL-M20 and ACR-D20 facilitated significantly more removal than Intersleek controls and were determined to be statistically equivalent to the PCL-PU control (ANOVA p < 0.0001). Although no significant trends in the biofilm removal data were observed based on the type of polyol and functionalized PDMS, the compositions prepared with 20 wt% PDMS consistently with the d-PDMS, regardless of the type of polyol utilized. The IS 900 coating also induced a high degree of biofilm retraction and achieved the lowest surface coverage (37%) among the coatings evaluated. Compositions PCL-M20, ACR-M10 and ACR-M20 were shown to be statistically equivalent to the IS 900 standard (ANOVA p < 0.0001).
The amount of C. lytica biofilm removal from the coating surfaces by water jetting at 81 kPa for 5 s is shown in Figure 2B (representative images in Figure S1B). Similar to the biofilm retraction analysis, the compositions based on m-PDMS facilitated the highest degree of biofilm removal (80-90%) and were statistically equivalent to the IS 900 standard (ANOVA p < 0.0001), which achieved 95% biofilm removal. In contrast, the IS 700 standard only enabled 65% of the retained biofilm to be removed and was not significantly different to the performance of the compositions based on d-PDMS (50-70% removal). The influence of the type of polyol (PCL or ACR) and the amount of PDMS incorporated into the coating (10 or 20 wt%) on C. lytica biofilm removal was shown not to be significant, except for the ACR-D series where statistically more for sporeling removal based on the compositional variations in polyol type (PCL and ACR) or PDMS content (10 and 20 wt%) (ANOVA p < 0.0001). The coatings derived from m-PDMS achieved significantly more removal than IS 700 (30%) and their performance was comparable to IS 900 (51%). Similar to the d-PDMS based compositions, sporeling adhesion was extremely tenacious to the PCL-PU control (6% removal), while the T2 silicone elastomer standard was comparable to IS900.

Adhesion of reattached adult barnacles
The adhesion of reattached adult barnacles is shown in Figure 3B, where the removal of nine barnacles was attempted for each coating. The entire set of siloxanepolyurethane coatings exhibited higher average adhesion values (0.21-0.39 MPa) than the IS 700 (0.11 MPa) and IS 900 (0.05 MPa) commercial FR coatings and the T2 silicone standard (0.18 MPa). Although the average adhesion value of barnacles reattached to the PU-PCL was shown to be similar to the siloxane-polyurethane coatings, a higher percentage of broken barnacles was observed on this surface during force gauge removal measurements (89%). In comparison, the siloxane-polyurethane coatings showed a considerably lower percentage of broken barnacles (11-33%), with the exception of coating PCL-D20 (67%). Coating PCL-M20 resulted in the lowest mean adhesion value (0.21 MPa) among the siloxane-polyurethane coatings and was determined to be significantly lower than PCL-D10 (0.39 MPa), PCL-D20 (0.38 MPa), ACR-M10 (0.34 MPa) and ACR-D20 (0.36 MPa) (ANOVA p < 0.0001). It is important to note that PCL-M20 was also shown to be statistically comparable to the IS 700 and T2 silicone standards, but was not statistically equivalent to achieved a greater amount of biofilm removal (24-26%) than their counterparts prepared with 10 wt% PDMS. Figure 2D displays the amount of N. incerta cell removal from the coating surfaces after water jetting at 81 kPa for 10 s. The PCL-PU control exhibited the highest amount of N. incerta removal among the coatings evaluated, which is typically observed for this type of coating system with regard to diatom adhesion studies in the laboratory (Mieszkin et al. 2012;Bodkhe et al. 2015). The coatings prepared with m-PDMS showed a higher amount of diatom removal (50-60%) than the coatings prepared with d-PDMS (28-45%) and the commercial FR control coatings (35-46%). It is also evident that the compositions prepared with d-PDMS and PCL polyol facilitated a higher amount of cell removal when compared to the analogous coatings prepared with ACR polyol, although this performance enhancement was only significant with respect to ACR-D10 (ANOVA p < 0.0001). The variation in PDMS content, however, was not shown to have a significant effect on N. incerta cell removal for either type of polyol (ANOVA p < 0.0001).

Adhesion strength of sporelings of U. linza
The removal of U. linza sporelings (young plants) from the coatings after water jetting at 111 kPa is shown in Figure  3A. In general, the amount of removal from the surface of the coatings prepared with d-PDMS was poor (0-17%) when compared to the coatings derived from m-PDMS (44-55%), while none of the attached sporeling biomass could be removed from PCL-D20 and ACR-D10 at this pressure. There were no significant differences observed and S3. In general, there was good agreement among the seven different assays, with the C. lytica biofilm retraction assay exhibiting the strongest degree of correlation to the other testing methods. In particular, this assay showed a high (r s = 0.70−1.0) to moderate (r s = 0.50−0.69) correlation with five of the other six laboratory testing methods, the highest being the removal of sporelings of U. linza (r s = 0.90; p = 0.0004) and C. lytica biofilm (r s = 0.88; p = 0.0009). These two screening methods also correlated strongly with each other (r s = 0.85; p = 0.0017) and to the N. incerta cell removal assay (U. linza r s = 0.76, p = 0.0104; C. lytica r s = 0.66, p = 0.0376), which was also shown to be significantly correlated to C. lytica biofilm retraction (r s = 0.76; p = 0.0113) and PB adhesion (r s = 0.78; p = 0.0075). In contrast, the H. pacifica biofilm removal and barnacle adhesion assays did not significantly correlate to any of the six other FR screening assays. However, the performance rankings for the barnacle adhesion assay were shown to agree to some extent with those from obtained with the C. lytica biofilm retraction assay (r s = 0.62), although the model was not statistically significant (p = 0.0563). This was not the case for the characterization of H. pacifica biofilm removal, where the Spearman's rank correlation coefficient was less than 0.25 (ie no correlation) when compared against the six other laboratory screening assays.

Static ocean immersion -Cal Poly
The removal of slimes, soft fouling and hard fouling via water-jetting after immersion for one month at the Morro Bay site at Cal Poly is shown in Figure 4A, where the water-jet pressure for the complete removal of each class of fouling is displayed. For slime removal, the coatings prepared with m-PDMS required a lower water-jet pressure compared to coatings prepared with d-PDMS, and the required pressure was similar to that observed for IS 700 and IS 900. It is also apparent that the d-PDMS coatings containing 20% PDMS showed higher removal than the IS 900 (ANOVA p < 0.0001). No significant trends in barnacle adhesion were observed with regard to the influence of polyol type or PDMS type and content of the siloxanepolyurethane coatings (ANOVA p < 0.0001).

Performance rankings
To help illustrate the overall similarities/differences in FR performance among the siloxane-polyurethane coatings, as well their relative performance to the commercial controls, a performance ranking (1-10; 1 = best, 10 = worst) was assigned to each coating as a function of each laboratory screening assay using the 'mean rank score' value derived from a non-parametric, one-way ANOVA (Kruskal-Wallis (rank sums) test; α = 0.05) ( Table 2). The top performing coating was the siloxane-polyurethane composition comprising 20 wt% m-PDMS and the PCL polyol (PCL-M20), with an average performance rank among all the assays of 2.6. IS 900 was determined to be the second best performing coating with an average performance rank of 3.5. Both PCL-M20 and IS 900 were identified as exhibiting the best performance in three out of the seven laboratory screening assays; pseudobarnacle adhesion, H. pacifica biofilm removal and N. incerta cell removal for PCL-M20 and C. lytica biofilm retraction, C. lytica biofilm removal and barnacle adhesion for IS 900. The three remaining compositions based on m-PDMS, namely PCL-M10, ACR-M10 and ACR-M20, showed the next best performance with average performance rankings of 3.6, 3.7 and 4.1, respectively. The average performance ranking for IS 700 and the d-PDMS series of siloxane-polyurethane coatings, on the other hand, were considerably worse (> 6.5).

Spearman's rank correlations for laboratory screening assays
In order to investigate the relationships among the different laboratory-based FR screening assays, Spearman's rank correlation coefficients (r s ) were calculated using the performance rankings for each individual FR laboratory testing method (Table 2) and are provided in Tables 3 Table 2. Ranking of fR performance of coatings evaluated with the fR laboratory assays. 2.6 7.9 7.6 3.7 4.1 7.9 7.4 6.7 3.5 SD 0.9 1.6 1.6 1.9 3.0 1.3 2.1 2.7 2.5 2.8 of fouling is shown. After immersion for three months, coatings PCL-M10 and ACR-M20 were cleaned at 80 psi, which was lower than the pressure required to clean IS 700 (0.83 MPa (120 psi)), but was higher than that required to clean IS 900 (0.27 MPa (40 psi)). The coatings prepared with d-PDMS all retained some level of soft fouling after water-jetting at 1.65 MPa (240 psi) (5-20%). The same was observed after immersion for six months, where the removal of soft fouling from the coatings prepared with d-PDMS also required higher water-jet pressures. However, most of the soft fouling was removed from the coatings prepared with m-PDMS, although considerably higher pressures were required for these compositions when compared to the commercial FR coatings IS 700 and IS 900. The removal of hard fouling by water-jet at Cal Poly after immersion for one month is also shown in Figure  4A. The coatings prepared with m-PDMS exhibited more complete removal of hard fouling at lower water-jet pressures than coatings prepared with d-PDMS. Furthermore, the coatings prepared with m-PDMS and PCL polyol required lower water-jet pressures than the rest of the siloxane-polyurethane coatings to achieve 100% removal of hard fouling and composition PCL-M20 was cleaned at the same pressure as IS 700 and IS 900 (0.55 MPa (80 psi)). The removal of hard fouling after immersion for three and six months is shown in Figure 4D. In general, the coatings based on m-PDMS exhibited enhanced FR properties when compared to the analogous coatings comprising d-PDMS. In particular, substantially more of the hard fouling remained on the PCL-D subset of coatings than the PCL-M compositions after water-jetting at the highest pressure (1.65 MPa (240 psi)) for both time points. None of the siloxane-polyurethane coatings, however, performed as well as the IS 700 and IS 900 commercial FR controls, which released all of the accumulated hard fouling at pressures lower than 1.65 MPa (240 psi).
The adhesion strength measurements and the percentage of broken barnacles recorded from force gauge measurements collected in the field at Cal Poly are analogous coatings containing 10% PDMS. The removal of slime after immersion for three and six months is shown in Figure 4B, where the remaining slime after water-jetting at 240 psi is reported. Since slime remained on all of the panels, except the copper control, it is evident that the FR performance was not as good as was observed after immersion for one month. With regard to immersion for three months, the siloxane-polyurethane coatings retained a greater amount of slime on the panels compared to IS 700 and IS 900. However, all the siloxane-polyurethanes, except PCL-M10, showed substantially more removal than IS 700 and their performance was comparable to IS 900 after immersion for six months. In contrast to the one-month immersion data, the m-PDMS coatings performed similarly to the d-PDMS coatings after immersion for three and six months, where PCL-D20 exhibited a relatively low amount of remaining slime (5%) following water-jetting at immersion for six months. In general, the siloxane-polyurethane coatings showed less slime remaining after water-jetting following immersion for six months compared to immersion for three months. The opposite behavior was observed for IS 700 and IS 900, where more slime remained following water-jetting after immersion for six months compared to immersion for three months.
The removal of soft fouling by water-jet after immersion for one month is shown in Figure 4A. Higher waterjet pressures were required for the removal of fouling from the coatings prepared with d-PDMS, compared to analogous coatings prepared with m-PDMS. Fouling remained on three of four coatings prepared with d-PDMS after water-jetting at 1.65 MPa (240 psi), suggesting that soft fouling removal from these coatings was low. The complete removal of soft fouling from siloxane-polyurethane coatings prepared with m-PDMS was obtained at a waterjetting pressure comparable to the pressure required to achieve complete removal from IS 700 (0.55 MPa (80 psi)), but required twice the pressure when compared to IS900 (0.27 MPa (40 psi)). The removal of soft fouling after immersion for three and six months is shown in Figure  4C, where the water jet pressure for complete removal  Figure 4E. It is clear from these dataset that the siloxane-polyurethane coatings that were prepared with m-PDMS (0.08-0.24 MPa; 0-47% breakage) exhibited superior FR properties when compared to the analogous . Results from field testing at Cal Poly and fiT. Water jetting at Cal Poly was performed on single replicate panels. Barnacle push-off adhesion at Cal Poly was for adult barnacles observed on the panels up to six months of immersion and the data labels above the black bars represent the number of barnacles included in the measurement. Water-jetting at fiT was performed on four replicate panels and the data labels represent the number of panels which were completely cleaned with the reported water-jet pressure. An asterisk indicates the panel was not cleaned at the maximum pressure of 1.38 MPa (200 psi). fiT barnacle adhesion was performed after immersion for 76 days and the data labels above the black bars represent the number of barnacles tested.
prepared with 10 wt% PDMS (0.08-0.13 MPa). Coatings PCL-M20, PCL-D20, ACR-M10, and ACR-M20 resulted in barnacle adhesion values that were comparable to IS 700. Interestingly, the barnacles attached to IS 900 became dislodged from the coating surface during the removal of the panels from the water to conduct force gauge measurements, indicating that the barnacles were only weakly adhered.

Static ocean immersion -NUS
Photos of panels immersed at The Republic of Singapore Yacht Club static immersion site for three months are shown in Figure 5, before and after soft sponging to remove fouling. Following immersion for three months, coatings prepared with d-PDMS did not clean upon soft sponging, and when barnacle removal was attempted, their shells/baseplates either broke during testing or the adhesion forces were in excess of 19.6 N. In general, the coatings prepared with m-PDMS cleaned better than the coatings prepared with d-PDMS when soft sponging was employed. However, coatings prepared with ACR polyol did not clean as well as those prepared with PCL polyol. This can be observed by examining the before and after cleaning images for PCL-M10 and ACR-M10, where ACR-M10 retained a substantial amount of fouling after soft sponging and PCL-M10 was almost completely cleaned. A similar trend was observed between PCL-M20 and ACR-M20, although to a lesser extent. Figure  6 displays the images of the coatings for which immersion and soft sponging was continued beyond three months due to their promising performance. IS 700 and IS 900 showed good cleaning ability, with almost complete removal of fouling with soft sponging. ACR-M20 showed fouling remaining after soft sponging, while PCL-M20 appeared to be cleaned of most fouling. The same was true for PCL-M10, where a majority of the fouling was removed with sponging, with the exception of barnacles, which were present in greater numbers than that of PCL-M20.

Performance rankings
To help illustrate the overall similarities/differences in FR performance among the siloxane-polyurethane coatings, as well their relative performance to the commercial controls, a performance ranking (1-10; 1 = best, 10 = worst) was assigned to each coating as a function of each field testing assessment conducted at Cal Poly and FIT using the 'mean rank score' value derived from a nonparametric, one-way ANOVA (Kruskal-Wallis (rank sums) test; α = 0.05) ( Table 4). Field testing assessments at NUS were excluded from performance rankings as they provided qualitative data only. The commercial standards (IS700, IS900) were shown to be the top two performing coatings with an average performance ranking of 2.5 and 3.7 for IS 900 and IS 700, respectively. Furthermore, < 0.12 MPa and released all attached barnacles during force gauge measurements without incurring visible damage to their shells or baseplates (ie 0% breakage). Additionally, both of these siloxane-polyurethane compositions performed comparably to the IS 700 (0.13 MPa; 14% breakage) and IS 900 (0.18 MPa; 0% breakage) commercial FR standards. Coating PCL-M10 was also shown to be comparable to the commercial Intersleek standards, exhibiting both a low average adhesion value (0.10 MPa) and percentage of broken barnacles (7%), while the analogous composition based on the ACR polyol (0.24 MPa; 47% breakage) was not nearly as effective.

Static ocean immersion -FIT
After immersion for 24 days at FIT's Indian River Lagoon static immersion site, the fouling consisted mostly of slime and soft fouling with some barnacles. The water-jetting pressure to remove slime fouling from a 2.5-3.8-cm diameter area on four panels is shown in Figure 4F, where the data labels represent the number of panels which were cleaned completely for each sample. It is readily apparent that the siloxane-polyurethane coatings based on the m-PDMS exhibited much better removal properties with regard to slime fouling than the coatings prepared with d-PDMS. Specifically, all four replicate panels enabled the slime fouling to be completely removed from the surfaces of compositions PCL-M10, PCL-M20 and ACR-M20 at a water-jet pressure below 1.38 MPa (200 psi).
In contrast, only one of the four replicates of ACR-D10 achieved complete slime removal at the maximum pressure utilized (200 psi), while none of the replicate panels could be completely cleaned for PCL-D10, PCL-D20 and ACR-D20. When comparing the performance of the siloxane-polyurethanes coatings to the commercial FR controls, compositions PCL-M10, PCL-M20 and ACR-M20 outperformed both IS 700 and IS 900, which required higher water-jet pressures to fully remove the attached slime fouling from only two and three of the four replicate panels, respectively. After immersion for 76 days the fouling consisted of barnacles, tubeworms, soft fouling and biofilm. Adult barnacles which were attached at this time (~5 mm in diameter) were subjected to force gauge measurements in shear and the results are shown in Figure 4F. One of the most obvious results was that barnacles could not be removed from ACR-D10 and ACR-D20 without incurring significant breakage to the shell and/or baseplate. The remainder of the siloxane-polyurethane coatings released barnacles intact, with average adhesion values ranging from 0.06 to 0.13 MPa. In general terms, barnacles attached to the compositions based on 20 wt% PDMS (0.06-0.08 MPa), regardless of the type of polyol or functionalized PDMS, were more easily removed than their counterparts series of siloxane-polyurethane coatings were shown to be considerably worse (> 6.5) than the m-PDMS series and the commercial FR standards.

Spearman's rank correlations for field testing assessments
In order to investigate the relationships among the different field testing assessments conducted at Cal Poly and FIT, Spearman's rank correlation coefficients (r s ) were calculated using the performance rankings for each individual field testing assessment (Table 4) and are shown in Tables 5 and S4. In general, there was good agreement among the various field testing assessments, with all but the FIT barnacle adhesion and Cal Poly slime adhesion after immersion for three and six months showing a moderate to high degree of correlation (r s ≥ 0.50) with at least six other field testing assessments. In this regard, the Cal Poly hard fouling adhesion after immersion for one and six months showed a high degree of correlation (r s ≥ 0.70) to eight and six of the 11 field assessment tests, respectively. The three-month Cal Poly soft fouling adhesion showed a high degree of correlation to six other field assessments, while the one-month Cal Poly slime and soft fouling adhesion were highly correlated to five. The Cal Poly barnacle adhesion, six-month soft fouling adhesion and FIT slime adhesion assessments were not as effective, achieving only three or four Spearman's rank correlation coefficient values ≥ 0.70 when compared to the 11 other field testing assessments. The three remaining field assessments, namely, three-and six-month Cal Poly slime adhesion and FIT barnacle adhesion, were shown to be highly correlated to either one or none of the other field testing assessments.
the IS 900 standard received the top ranking in eight of the 12 field assessments, while the IS 700 received the top ranking in five. The best-performing coating from the siloxane-polyurethane coatings was the composition comprising 20 wt% m-PDMS and the PCL polyol (PCL-M20), which scored an average ranking of 3.8 and was shown to be the best-performing coating in two of the field testing assessments. The three remaining compositions based on m-PDMS, namely ACR-M20, PCL-M10 and ACR-M10, showed the next best performance with average performance rankings of 4.4, 4.9 and 5.5, respectively. The average performance ranking for the d-PDMS  C. lytica biofilms was highly correlated to four field assessments, while the N. incerta cell removal and pseudobarnacle adhesion assays were highly correlated to only two and one field testing assessment, respectively. Interestingly, the H. pacifica biofilm removal assay showed no correlation to 10 of the 12 field testing assessments, with only a low degree of correlation (r s = 0.30−0.49) established between this laboratory screening assay and three-and six-month Cal Poly slime adhesion assessments in the field.

Overall coating performance
A comparison of the average performance rankings for each coating in the laboratory screening assays (Table  2) and field testing assessments (Table 4) is provided in Figure 7. In both the laboratory assays and the field assessments, coatings PCL-M10, PCL-M20, ACR-M20 and IS 700 were shown to be statistically equivalent to the IS 900 commercial FR standard. Coating ACR-M10 was also determined to be statistically equivalent to IS 900

Spearman's rank correlations
In order to investigate the relationships between the laboratory screening assays and the field testing assessments conducted at Cal Poly and FIT, Spearman's rank correlation coefficients (r s ) were calculated using the performance rankings for each laboratory assay and field assessment (Tables 2 and 4) and are provided in Tables 6 and S5. It is evident that the C. lytica biofilm retraction assay was the most effective laboratory screening method with regard to the agreement with static ocean immersion testing, showing a high degree of correlation (r s ≥ 0.70, p < 0.05) with seven of the 12 field testing assessments. The reattached barnacle adhesion and U. linza removal assays also showed good agreement with the results from ocean immersion testing, with both laboratory screening methods exhibiting a high degree of correlation to five of the field testing assessments. The water-jetting removal of    For FR applications, these laboratory assessments may include measurements of surface energy, pseudobarnacle adhesion and methods to assess the adhesion strength of representative organisms to the coating surface. It is important, however, to determine the relevance of the laboratory assessments to the FR behavior of the coatings in the more complex marine environment.
In this study, one of the most obvious conclusions that can be made when conducting a cursory examination of the results is that the siloxane-polyurethane coatings prepared from m-PDMS were considerably more effective at mitigating biofouling adhesion than those generated from d-PDMS. This marked difference is well illustrated in the average performance rankings for each siloxanepolyurethane coating provided in Table 2 (laboratory assays), Table 4 (field assessments) and Figure 7. The average performance ranking for the four compositions comprising m-PDMS ranged from 2.6 to 4.1 in the laboratory and 3.8-4.9 in the field. In comparison, the four compositions composed of d-PDMS possessed an average performance ranking that ranged from 7.4 to 7.9 in the laboratory and 6.9-8.0 in the field. Furthermore, three of the four compositions based on m-PDMS (PCL-M10, PCL-M20 and ACR-M20) had an average performance ranking that was statistically equivalent to the IS 900 commercial FR coating system in both the laboratory assays and field assessments, with PCL-M20 identified as the top performing siloxane-polyurethane coating in both arenas of testing. Although the results from static ocean immersion testing at NUS were qualitative, it is clear that these three m-PDMS coatings were cleaned of fouling easier by sponging than the rest of the siloxane-polyurethanes ( Figure 5) and that PCL-M20 was the most efficient in this regard ( Figure 6). Neither the laboratory nor the field testing assessments, however, were able to distinguish significant differences in FR performance as a function of polyol type (PCL or ACR) or the amount of functionalized PDMS (10 or 20 wt%) incorporated in the laboratory screening assays. In contrast, none of the siloxane-polyurethane coatings derived from d-PDMS were shown to be statistically equivalent to the IS 900 in either the laboratory assays or the field assessments. The three best-performing coatings in the laboratory assays were PCL-M20, IS 900 and PCL-M10, which possessed an average performance ranking of 2.6, 3.5 and 3.6, respectively. Likewise, the IS 900 and PCL-M20 were also identified as one of the top three performing candidates in the field testing assessments, with an average performance ranking of 2.5 and 3.8, respectively. However, PCL-M10 did not perform as well in the field (average rank 4.9), but was still identified as one of top five performing coatings. Interestingly, the IS 700 commercial standard was shown to be much more effective in mitigating fouling adhesion in the field assessments (average rank of 3.7) than in the laboratory screening assays (average rank 6.7).

Discussion
New coatings technologies designed to mitigate biofouling need to be assessed in the laboratory prior to carrying out Table 6. Spearman's rank correlation coefficients (r s ) for the comparisons between individual laboratory assays and individual field assessments. CP = Cal-Poly, fiT = florida institute of Technology, 1 = one month immersion, 3 = three months immersion, 6 = six months immersion, Barn = barnacle, PB = pseudobarnacle. A version of this  Figure 7. Average performance rankings for the siloxanepolyurethane and commercial standards evaluated with the laboratory screening assays and field assessments. Coatings designated with an asterisk are statistically equivalent to the iS 900 commercial coating (one-way AnoVA, α = 0.05).
the field, although four out of five of these correlations were established with the early onset of fouling only (ie one month). Interestingly, the C. lytica biofilm retraction and U. linza adhesion assays had been previously demonstrated to correlate well with one another in a study that investigated the FR properties of an early generation of siloxane-polyurethanes based exclusively on d-PDMS (Ribeiro et al. 2008). The reattached barnacle adhesion assay was also highly correlated to five of the field testing assessments and included all three durations of exposure, but was limited to soft and hard fouling only. It is interesting to note that the laboratory reattachment assay did not correlate well with barnacle adhesion measurements obtained at Cal Poly (r s =0.51, p = 0.1315) and FIT (r s =0.43, p = 0.2098). However, these two field testing sites also showed a low degree of correlation between their respective barnacle adhesion assessments for this set of coatings (r s =0.37, p = 0.2953). Although an explanation for these seemingly disparate results among the three testing sites is not readily apparent, a number of factors could have influenced the FR performance of the coatings with respect to the mitigation of barnacle adhesion, including variations in the temporal and spatial aspects of the fouling community (field assessments), barnacle species (B. crenatus, Cal Poly; B. eburneus, FIT; A. amphitrite, NDSU), water temperature, salinity, pH and dissolved organic content Swain et al. 2000;Wood et al. 2000;Yebra et al. 2004;Holm et al. 2008;Johnston 2010). It has also been shown that hard fouling adhesion varies among different fouling organisms (Kavanagh et al. 2001). The results of the laboratory reattachment assay, however, appear to agree quite well with the soft sponging assessments conducted at NUS. Specifically, a close inspection of the images taken after sponging ( Figure 6; see middle/interior region of panels and ignore nonsponged edges) revealed that PCL-M10, PCL-M20, IS 700 and IS 900 were the most efficient at releasing barnacles, with PCL-M20 shown to be the most effective of the siloxane-polyurethane coatings. These results obtained at NUS were entirely consistent with the results from the laboratory reattachment assay, which also ranked these coatings as the top four performers and identified PCL-M20 as the most effective siloxane-polyurethane composition. Finally, the C. lytica biofilm removal assay was highly correlated to four field assessments which included all three classes of fouling after immersion for one and three months. The remaining laboratory screening assays, namely PB adhesion, N. incerta cell removal and H. pacifica biofilm removal, were not shown to be as effective as a 'predictive indicator' of overall FR performance observed for the siloxane-polyurethane and standard coatings in the static ocean immersion trials. In particular, the H. pacifica into the siloxane-polyurethane coating system, although in some instances, the individual laboratory assays were able to discern statistical differences in performance based on these two formulation variables (eg C. lytica biofilm retraction; Figure 2A).
The possible reason for the better performance of the coatings based on m-PDMS over the coatings made using d-PDMS is believed to originate from the nature of the chemical binding of the PDMS into the coating system. Since d-PDMS has two end-chain functional groups, the siloxane is anchored at both chain ends into the polyurethane coating. In contrast, since m-PDMS has the functional group at only a single chain end, the siloxane is anchored at only one chain end and therefore the siloxane chain has a greater degree of freedom on the coating surface. In previous work, it was found that siloxane-polyurethane coatings made using m-PDMS had good FR properties over a wide range of siloxane molecular weights and amounts (Sommer et al. 2010). A study of the thickness of the siloxane layer of these coatings using Rutherford backscattering indicated that the coatings based on m-PDMS had a thicker siloxane layer than those based on d-PDMS (Siripirom 2012). The greater siloxane thickness is likely to be responsible for the easier removal of biofouling from these surfaces.
When taking into account each laboratory assay on its own, the C. lytica biofilm retraction assay was the most effective screening method for predicting the FR performance results obtained in static ocean immersion trials. As observed in Table 6, this screening technique was highly correlated to seven of the 12 metrics assessed in the field and included all three classes of fouling (ie slime, soft and hard/barnacles) and all three immersion times (ie one, three and six months). This relatively strong and broad degree of correlation to the field is surprising when it is considered that the biofilm retraction assay was the most 'pedestrian' of the laboratory screening methods, relying simply on the capillary forces exerted on the attached biofilm during drying rather than shear-forces applied with a water-jet or force gauge to elucidate differences in the easy-release properties among the coatings (Stafslien et al. 2007a). Furthermore, these results appear to corroborate a previous study reported by the authors, where the C. lytica biofilm retraction assay was shown to strongly correlate to both the overall fouling rating and barnacle adhesion measurements collected at FIT for a set of polysiloxane FR coatings (Stafslien et al. 2007a).
The next subset of laboratory screening techniques that correlated relatively well to field testing was the adhesion strength of sporelings of U. linza, reattached barnacle adhesion and C. lytica biofilm removal assays. As with the C. lytica biofilm retraction assay, U. linza removal was also highly correlated to all classes of fouling assessed in water-jet slime removal at both Cal Poly and FIT, and as indicated above, clearly distinguished the m-PDMS subset of coatings as possessing superior FR properties than the analogous compositions comprising d-PDMS, as did each of the water-jet removal assessments at Cal Poly and FIT after immersion for one month.
In contrast to the correlations established with the early onset of fouling, the water-jet based laboratory assays did not correlate as well to the fouling removal observed after immersion for three and six months at Cal Poly. This was particularly evident for slime and hard fouling, where each water-jet based laboratory assay showed either a low degree or no correlation at all after immersion for three and six months. The C. lytica biofilm and U. linza sporeling removal assays, however, were shown to highly correlate to soft fouling removal after immersion for three months, but not after immersion for six months. It is important to point out that the slime fouling removal results after immersion for one month at Cal Poly did not agree very well with the slime fouling removal results obtained after immersion for three and six months, with r s values of −0.06 and −0.58, respectively. Similarly, the soft fouling removal after immersion for six months at Cal Poly showed only a moderate to low degree of correlation to the soft fouling removal observed after immersion for one and three months. As might then be expected, the performance ranking for each coating was also variable and dependent upon the immersion time. For example, siloxane-polyurethane composition PCL-M10 exhibited an average ranking of 1.0, 7.0 and 9.0 for slime fouling and 3.5, 2.5 and 7.5 for soft fouling after immersion for one, three and six months, respectively. Thus, it is not surprising that the short-term water-jet based laboratory assays did not correlate particularly well with the later stages of slime and soft fouling removal at the Cal Poly field testing site due to these temporal variations in coating performance.
As opposed to slime and soft fouling removal, which correlated relatively well to both C. lytica biofilm and U. linza sporeling removal after immersion for one and three months, the water-jet based laboratory assays were not as effective an indicator of hard fouling removal, regardless of the immersion time, with U. linza sporeling removal correlated to one month hard fouling removal being the only exception (r s = 0.70, p = 0.0251). The hard fouling removal results, however, were much more consistent over the six-month immersion at Cal Poly than observed for slime and soft fouling, with all three time points shown to correlate exceptionally well with one another (r s ≥ 0.90). Likewise, the trends in FR performance based on coating type were also more consistent for this class of fouling, with the siloxane-polyurethane composition PCL-M10, which showed considerable biofilm removal assay showed a low degree of correlation to two of the field tests while exhibiting no correlation at all to the 10 remaining assessments. It is interesting to note, however, that this bacterium exhibited tenacious biofilm adhesion to both of the commercial FR coatings, with a performance ranking of 10.0 and 7.0 assigned to IS 700 and IS 900, respectively. These results were somewhat consistent with the results of the slime adhesion assessment after one month at FIT (6.0 for IS 700; 5.0 for IS 900) and six months at Cal Poly (10.0 for IS 700; 5.5 for IS 900), which suggests that the H. pacifica biofilm removal assay, although not suitable for ascertaining broadspectrum FR performance, may be a good indicator of FR performance with respect to aggressively-adhering slime communities in the field. Similarly, N. incerta cell adhesion was also shown to be quite tenacious to the Intersleek coating systems, exhibiting a performance ranking of 8.0 and 5.0 for IS 700 and IS 900, respectively. However, this laboratory assay was more effective than the H. pacifica biofilm removal assessment as it was highly correlated to two field assessments and moderately correlated to four. The N. incerta cell removal assay also clearly distinguished the m-PDMS coatings as possessing superior release properties when compared to the d-PDMS coatings, which was not the case for H. pacifica; especially for the ACR subset of compositions. Likewise, the results of the PB adhesion assay showed that the compositions based on m-PDMS were more effective at attenuating the adhesion strength of pseudobarnacles when compared to their d-PDMS counterparts. This laboratory screening technique was less effective, however, at predicting broad-spectrum FR performance in the field, showing only a single high and moderate degree of correlation among the 12 static ocean immersion assessments. Previous research by Swain (1997) has shown that there is not always a direct correlation between pseudobarnacle and live barnacle adhesion strength.
It is worth noting that, with the exception of H. pacifica biofilms, the water-jet screening assays in the laboratory correlated very well with water-jet removal assessments in the field at both Cal Poly and FIT for early onset of fouling (ie immersion for one month). For example, the C. lytica biofilm, N. incerta cell and U. linza sporeling removal assays all showed a moderate to high degree of correlation to the water-jet removal of all three classes of fouling at Cal Poly and slime at FIT after immersion for one month. The most effective was the U. linza sporeling removal assay, which was both highly and significantly correlated to all four of the one-month water-jet based field assessments, while the C. lytica biofilm and N. incerta cell removal assays were highly and significantly correlated to two of them. However, all three of these water-jet based laboratory assays were significantly correlated to one-month predicted the broad-spectrum release properties of each coating towards slime, soft and hard fouling communities. The most effective assay based on these metrics was the C. lytica biofilm retraction assay (Tier 3), followed by U. linza sporeling removal, reattached barnacle adhesion and C. lytica biofilm removal (Tier 2) and finally, the N. incerta cell removal, pseudobarnacle adhesion and H. pacifica biofilm removal assays (Tier 1). However, these tiers can be further refined when a Spearman's rank correlation is carried out for each individual laboratory screening assay compared to the average performance ranking for all of the field assessments (Table 7). From this analysis, the barnacle reattachment technique would be re-categorized from a Tier 2 to a Tier 3 screening assay and the N. incerta cell removal assay would be re-categorized from Tier 1 to Tier 2 ( Figure 8). Based on this tiered categorization/ classification system, five of the seven laboratory screening assays were shown to be highly effective as predictive indicators of the overall FR performance of siloxane-polyurethane coatings determined at three different field testing sites over the course of static ocean immersion for one to six months. Additionally, three of these laboratory assays, namely, C. lytica biofilm retraction (Stafslien et al. 2007a), reattached barnacle adhesion (Rittschof et al. 2008;Stafslien et al. 2012) and U. linza sporeling removal (Zhang et al. 2013), have also been shown to correlate well to field testing for other types of AF/FR coating systems, which further demonstrates their capacity to be used as a preliminary screening tool to assess a wide range of AF/ FR coating technologies.
From a materials/coatings development standpoint, it would be extremely beneficial to subject new FR technologies to the full battery of laboratory screening assays detailed in this study to glean a thorough and comprehensive understanding of their easy-release properties. This information could then be used to aid in the determination of which candidate technologies merit the expenditure of additional time, money and other valuable resources required to further assess their performance profiles in real-world environments through various types of ocean immersion testing procedures. However, this allencompassing laboratory screening approach may not be practical for all types of emerging coating technologies, such as those that are subject to limited availability and/ or high cost of the materials used in the early phases of their development. In these instances, a first approximation of their broad-spectrum FR performance potential could be ascertained by employing one or both of the Tier 3 laboratory screening techniques depicted in Figure 8, namely, reattached barnacle adhesion and C. lytica biofilm retraction assays. This would require the preparation of only two replicates of 2.5 × 7.6 cm specimens and four replicates of 15-mm disks (ie well bottom area of 24-well variation for slime and soft fouling, achieving a performance ranking of 4.0, 5.5 and 4.5 after immersion for one, three and six months. These observations suggest that temporal variations in fouling at the Cal Poly field-testing site were not a primary factor in the lower degree of correlation observed between the water-jet based laboratory assays and hard fouling removal. A more likely and logical explanation would be that these laboratory screening assays were based on slime (bacteria and diatoms) and soft fouling (U. linza) organisms, and thus would not be expected to correlate well to hard fouling removal measurements collected in the field due to differences in their adhesive strategies (Petrone 2013). However, it should be noted here that the results of the C. lytica biofilm retraction assay were highly and significantly correlated to hard fouling removal at all three immersion time points. This assay also showed a moderate degree of correlation with the barnacle adhesion measurements in the laboratory reattachment assay (r s = 0.62, p ≤ 0.0563) and at both the Cal Poly (r s = 0.63, p = 0.0493) and FIT (r s = 0.58, p = 0.0805). Interestingly, the biofilm retraction assay has been shown in a previous study to agree very well with barnacle adhesion measurements made on polysiloxane-based FR coatings at the FIT field-testing site (Stafslien et al. 2007a). Thus, this assay appears to be a simple and effective method for ascertaining the preliminary easy-release properties of siloxane-based FR coating systems with respect to barnacle fouling.
Although the barnacle reattachment assay did not correlate very well to barnacle adhesion measurements collected at Cal Poly and FIT, as discussed above, it was shown to correlate highly with the hard-fouling removal data collected at Cal Poly for all three immersion time points (r s ≥ 0.88). Again, this may not be too surprising or unexpected since barnacles are a representative organism of this class of fouling. It is worth reiterating here that the barnacle reattachment assay did exhibit good agreement with the qualitative assessments of hard fouling/barnacle removal at NUS using a soft sponging technique. These observations suggest that the two-week laboratory reattachment assay may be a relatively quick and powerful screening technique to gauge a coatings FR properties with regard to both barnacle and hard-fouling adhesion in the field.
The preceding sections of this discussion delineated the primary strengths and weaknesses of the individual laboratory screening assays in the context of their relevance to the FR performance of coatings gleaned from static ocean immersion trials at three field testing sites, as a function of performance rankings and the Spearman's rank correlation coefficients derived from them. In this regard, the seven laboratory techniques were separated into three general categories or 'tiers' based on how well they a subset of laboratory assays based solely on macrofouling species, namely, reattached barnacle adhesion and U. linza sporeling removal, to initially screen the candidate technologies. Those compositions showing good release properties for adult barnacles and microalgal spores could then be subjected to one or more microfouling-based assays to narrow down and identify only the most promising or 'lead' candidates for field testing trials. A different approach may be desirable for vetting technologies intended for merchant vessels, which typically spend the majority of their time in transit rather than stationed pier side. Since these vessels are under constant operation, the accumulation of larger organisms on their hull tends to be less severe, allowing microfouling slime communities to flourish (Hunsucker et al. 2014). As such, it may be more practical to subject candidate FR technologies being developed for merchant ships to preliminary performance assessments with a subset of slime fouling assays, namely, C. lytica biofilm retraction and removal and N. incerta cell removal assays. However, the choice of assays may also be influenced by other factors (ie regional fouling dynamics, vessel operational cycles) and thus the testing strategies outlined above are provided as a general framework from which to construct the most prudent testing scheme for the technology being investigated.
In the present study, the laboratory screening assays were run in parallel with static ocean immersion testing as a means of examining or establishing their predictive plate) for each candidate technology, a relatively modest amount of material and effort.
The selection of laboratory screening assays can also be dictated by the targeted application(s) of the particular FR technology under development. For example, coatings that are intended to prevent or mitigate fouling on US Navy vessels must contend with a relatively high degree of fouling pressure during lengthy stays in port. These prolonged periods of inactivity are highly conducive to the accumulation of both microfouling and macrofouling communities, where barnacles, tubeworms, bivalves and macroalgae are some of the most problematic organisms in the context of increasing a vessel's hydrodynamic drag. In this case, an investigator may choose to utilize Table 7. Spearman's rank correlation coefficients (r s ) and p-values for the comparisons between individual laboratory assay rankings and average field assessment rankings. PB Figure 8. illustration of the predictive capability of the laboratory-based fR screening assays to the broad-spectrum fR performance observed at fiT, Cal Poly and nuS static ocean immersion field testing sites.
resulted in marginal performance improvements over those comprising 10 wt% PDMS, with composition PCL-M20 displaying the best overall FR performance in both regimes of testing and comparable to the commercial FR standard coating systems from International Paint (ie IS 700 and IS 900). Of the laboratory screening techniques, the reattached barnacle adhesion and C. lytica biofilm retraction assays (Tier 3 assays) were shown to be the most predictive of FR performance in the field, exhibiting a high degree of correlation to seven and five of the 12 field assessments, respectively, and highly and significantly correlating to the average field assessment rankings. Both of these techniques correlated very well to short (ie one month) and longer-term performance observed towards soft and hard fouling in the field, but only to short-term performance against slimes at the Cal Poly testing site due to temporal variations at exposure for three and six months. The macroalgal (U. linza) sporeling, microalgal (N. incerta) cell removal and C. lytica assays (Tier 2) were also quite effective, but to a lesser extent than the Tier 3 assays, while the pseudobarnacle adhesion and H. pacifica biofilm assays (Tier 1) had limited or no predictive capability with respect to broad-spectrum field performance of the siloxane-polyurethane FR coatings.
relationship to the various field assessments. Thus, they were not leveraged prior to field testing trials to down-select the siloxane-polyurethane coating series to a smaller subset of compositions. However, it may be instructive to illustrate how the results of the laboratory screening assays would have functioned in this capacity, had they been used as a preliminary screening tool as delineated in the two preceding paragraphs. In this regard, the combination of laboratory characterization techniques listed in Tier 3 and Tier 2 of Figure 8 would have effectively eliminated the d-PDMS series from the set of siloxane-polyurethane coatings to be scaled up and evaluated in the field. The exclusion of the d-PDMS series from field testing would have substantially reduced the number of 10.16 cm x 20.32 cm (4" x 8") panels that needed to be prepared from 96 (eight compositions × four replicates × three field testing sites) down to 48 (four compositions). If material constraints were a concern, the utilization of the Tier 3 assays exclusively may have eliminated the ACR series completely, as the performance rankings from the barnacle reattachment assay, the most highly correlated laboratory technique to field testing, clearly distinguished the m-PDMS PCL compositions as the two best performing materials among the eight siloxane-polyurethane coatings. Down-selection in this manner would have further reduced the number of 10.16 cm x 20.32 cm (4" x 8") panels required for field testing from 48 down to 24 (two compositions × four replicates × three field testing sites), resulting in both significant savings in labor and material costs needed to produce the coated panels and the amount of time, money and resources required to conduct the various assessments at the Cal Poly, FIT and NUS static ocean immersion sites. This, then, is the real benefit of having a suite of laboratory-based screening assays that can provide some degree of predictive capability of realworld performance readily available to materials and coatings developers.

Conclusions
The combination of laboratory-based screening assays described in this study was shown to correlate very well to the broad-spectrum performance data collected for a series of eight siloxane-polysiloxane FR coatings at three different static ocean immersion testing sites.
In particular, several of the laboratory screening techniques were able to effectively 'predict' field testing performance as a function of compositional variation, where the coatings based on m-PDMS exhibited superior FR properties when compared to the analogous compositions based on d-PDMS. Furthermore, both the laboratory and the field assessments revealed that the siloxane-polyurethanes comprising 20 wt% PDMS