Version of Record: https://www.sciencedirect.

com/science/article/pii/S0956713519305158
Manuscript_874e713ec8e1545fc9651039ffa72e74

1 Use of 254 nm ultraviolet light for decontamination of fresh produce and wash water

3 Runze Huang1, Haiqiang Chen1*

4 1Department of Animal and Food Sciences, University of Delaware, Newark, DE 19716-2150,

5 USA

6 *Corresponding author: Tel.: +1-302-831-1045; Fax: +1-302-831-2822.

7 Email address: haiqiang@udel.edu (H. Chen).

© 2019 published by Elsevier. This manuscript is made available under the Elsevier user license
https://www.elsevier.com/open-access/userlicense/1.0/
9 Abstract

10 Fresh produce has been identified as the transmission vehicle for many foodborne illness

11 outbreaks every year. They could be contaminated by pathogens at any point within the farm-to-

12 fork food chain. Washing is often used to remove dirt and pathogens on fresh produce. Moreover,

13 an effective decontamination method is required for wash water to prevent cross-contamination.

14 Our group developed a water-assisted ultraviolet (UV) decontamination system (WUV) which

15 combined washing and UV treatments. We evaluated the influence of turbidity (0-100 NTU) and

16 chemical oxygen demand (COD; 0-2500 ppm) on UV inactivation of Salmonella in simulated

17 wash water with different produce extracts. In the worst condition tested (turbidity of 100 NTU

18 and COD of 2500 ppm), Salmonella in simulated wash water could be eliminated by UV (23-28

19 mW/cm2; wavelength peak at 254 nm) in 90 seconds (> 6 log CFU/mL reduction), indicating the

20 potential of UV for fresh produce wash water disinfection. Within the ranges tested, the

21 effectiveness of UV inactivation of Salmonella was significantly decreased with increasing COD,

22 but not with turbidity. In addition, ultraviolet lamps were installed in different locations (above

23 and/or in the washing tank).Three WUV configurations showed similar Salmonella reduction on

24 blueberries (1.8-2.0 log) and tomatoes (2.4-2.9 log). However, the Overhead and Dual WUV

25 configurations reduced significantly (P < 0.05) more Salmonella on lettuce than the Submersible

26 WUV configuration.

27 Keywords: Submersible ultraviolet; Salmonella; turbidity; chemical oxygen demand; fruits and

28 vegetables

29

2
30 1. Introduction
31 Fresh fruits and vegetable have been proven to be beneficial to human health, and many

32 countries have made great efforts to promote the intake of fresh produce (Rekhy & McConchie,

33 2014). The global production of fruits and vegetables maintains a steady increase over the last 30

34 years, reaching a total of 1.94 billion tons in 2016 (Food and Agriculture Organization (FAO),

35 2018). One undesirable side effect with this trend is the accompanying pathogen contamination

36 risk of fresh produce. Bennett et al. (2018) showed that there were 972 raw produce related

37 outbreaks in the U.S. from 1998 – 2013, and the percentage of raw produce related outbreaks out

38 of all foodborne outbreaks showed an increasing trend in the period. In early 2018, 210 people

39 across the U.S. were infected by Escherichia coli O157:H7 due to consumption of romaine

40 lettuce and 5 deaths were reported (Centers for Disease Control and Prevention, 2018). Later in

41 2018, another multistate foodborne illness outbreak of pathogenic E. coli in the U.S. was linked

42 to lettuce, causing at least 59 illnesses (CDC, 2018). The contaminated food also imposed

43 serious economic burdens. Hoffmann, Maculloch, & Batz (2015) estimated that the economic

44 loss caused by foodborne illnesses in the U.S. was > 15.5 billion dollars per year.

45 Currently, chlorine is the most commonly used disinfectant for fresh produce washing

46 due to its low cost and relatively good decontamination efficacy (Goodburn & Wallace, 2013; S.

47 Van Haute, Tryland, Veys, & Sampers, 2014). Unfortunately, this practice presents several

48 drawbacks: leaving disinfectant residues on produce items (Goodburn & Wallace, 2013) and

49 producing toxic by-products (Tudela et al., 2019). In addition, a large amount of chlorine need to

50 be added to wash water to maintain an adequate free chlorine level since it reacts very easily

51 with organic materials. Some European countries have banned the use of chlorine in wash water

52 for ready-to-eat products including fresh produce (Coroneo et al., 2017). On the other hand,

3
53 short-wave ultraviolet (UV) has been widely used for food decontamination. It is a physical

54 method and does not produce any toxic by-products (Fan, Huang, & Chen, 2017). Additionally,

55 the capital and operating costs of UV are low, make it suitable for industrial applications (Garg,

56 Abela, Tiwari, & Valdramidis, 2016).

57 Recently, our group combined washing and UV to develop a water-assisted UV (WUV)

58 system (Guo, Huang, & Chen, 2017; Huang, de Vries, & Chen, 2018). To perform the combined

59 treatment, swirling water was used to wash produce and a UV lamp was installed above the

60 water. The washing process eliminates the non-uniform exposure of common UV treatments, and

61 the UV treatment helps kill the pathogens eluted into water (Huang et al., 2018). However, due

62 to the poor penetration ability of UV, this WUV system faced two problems for the application

63 on fresh produce. First, soil, fertilizers, and exudates from produce could be introduced into wash

64 water, which would lower the water quality and weaken the antimicrobial effect of UV (Selma,

65 Allende, López-Gálvez, Conesa, & Gil, 2008). Second, fresh produce, especially leafy

66 vegetables, tends to float on the water surface and form a barrier against UV light, resulting in

67 insufficient exposure to UV for the underneath produce and water (Huang et al., 2018).

68 The poor penetration ability of UV has been demonstrated by many studies (Fan et al.,

69 2017). As for UV inactivation of pathogens in liquid, the efficacy is affected by many factors,

70 including the UV absorbing compounds, insoluble compounds, depth of liquid and intensity of

71 UV (Koutchma, 2009). Previous studies showed that UV could be used to disinfect fresh produce

72 wash water so that it could be reused, resulting in the reduction of water consumption (Ignat,

73 Manzocco, Bartolomeoli, Maifreni, & Nicoli, 2015). Various parameters were used to describe

74 the quality of produce wash water, including turbidity, chemical oxygen demand (COD), total

75 dissolved solids, ions concentration, etc. (Nou et al., 2011; Sam Van Haute, Sampers, Holvoet, &

4
76 Uyttendaele, 2013). Out of these parameters, COD and turbidity are two parameters that are

77 commonly used to describe the condition of fresh produce wash water (Sam van Haute, Sampers,

78 Jacxsens, & Uyttendaele, 2015). The typical COD level found in fresh produce wash water

79 ranged from 200 to 2000 ppm (Cossu, Ercan, Tikekar, & Nitin, 2016). Some higher COD levels

80 were found in wash water for onions (5461 ppm) (S. Van Haute et al., 2018) and potato (18000

81 ppm) (Sayed, El-Ezaby, & Groendijk, 2005). As for the turbidity of water for washing fresh

82 produce, it is usually below 100 NTU (Luo et al., 2012; Nou et al., 2011). In this study, we

83 prepared simulated wash water with different fresh produce extracts and determined the

84 influence of turbidity and COD on Salmonella survival in simulated wash water while being

85 treated by UV. The levels of turbidity and COD of wash water tested in this study was

86 comparable to the typical conditions.

87 To solve the problem of UV blockage caused by the floating leafy greens, one potential

88 way is to install UV lamps in water. In this way, the fresh produce could receive UV exposure

89 from both above and below, potentially resulting in more uniform exposure. Commercially

90 submersible UV lamps are used for disinfection of wastewater and have been shown to be

91 effective against many types of microorganisms (United States Environmental Protection

92 Agency, 1999). In this study, we installed UV lamps above the washing tank as well as in it.

93 This study aimed to: 1) evaluate the effect of turbidity and COD on the inactivation of

94 Salmonella in wash water by UV and 2) compare the efficacy of 3 different WUV configurations

95 on Salmonella inactivation on blueberry, tomato and lettuce shreds.

96 2. Materials and Methods

97 2.1. WUV configurations

5
 98 Three WUV configurations were used in this study. For the Overhead WUV

99 configuration as shown in Fig. 1, produce samples were washed by agitated water (agitated with

100 two stirring pads rotated at 150 RPM) in a stainless steel tank (53 cm (L) × 33 cm (W) × 15 cm

101 (H)) while being exposed to four UV lamps (wavelength peak at 254 nm; ~17 cm from the water

102 surface; 23-28 mW/cm2; Heraeus Noblelight, Buford, GA) mounted over the tank (Huang et al.,

103 2018). The second was the Submersible WUV configuration (Fig. 2). Samples were washed by

104 agitated water in a stainless steel tank while being exposed to a submersible UV lamp

105 (wavelength peak at 254 nm; 4 mW/cm2; Steril-Aire Hydro Submersible 16, Steril-Aire, Burbank,

106 CA) installed under water. The third was the Dual WUV configuration, in which the combination

107 of the Overhead and Submersible configurations was used. Samples were washed by agitated

108 water while being exposed to both the overhead UV lamps and the submersible UV lamp.

109 2.2. Salmonella inoculum preparation

110 The Salmonella inoculum preparation followed the method described by Huang et al.

111 (2018). Briefly, four nalidixic acid (100 μg/mL) and streptomycin (100 μg/mL) resistant strains

112 of Salmonella enterica (S. Heidelberg 45955, S. Montevideo 51, S. Newport H1073 and S.

113 Typhimurium14028; kindly provided by Dr. Vivian Wu from USDA) were grown at 35 °C for

114 24 hours to stationary phase individually. Two consecutive subcultures of each strain were

115 performed. The four Salmonella cultures were mixed and centrifuged at 4000 × g for 10 min.

116 The precipitate was resuspended in 0.1% peptone water and mixed. The concentration of the

117 Salmonella inoculum cocktail was ~109 CFU/mL.

118 2.3. Influence of turbidity and COD on inactivation of Salmonella in simulated wash water

119 by UV

6
120 Tap water (10 L) at 4 °C and 100 mL of Salmonella inoculum were added into the

121 washing tank of the Overhead WUV configuration, and the final concentration of Salmonella

122 was ~ 2 × 106 CFU/mL. Extracts of fresh produce were used to adjust the COD levels of wash

123 water and silicon dioxide suspension (SiO2 suspension) were used to adjust the turbidity levels

124 using the procedures described in Huang et al. (2018). Turbidity and COD were determined

125 using a turbidity meter (HI 93414; Hanna Instruments, Woonsocket, RI) and a COD meter (YSI

126 910; YSI, Yellow Spring, OH), respectively.

127 Four experiments were carried out to evaluate the effects of turbidity and COD on

128 Salmonella inactivation in wash water while being treated by UV. 1) Effect of COD (0 – 2500

129 ppm): adjusted using tomato extract (0 – 5%) with the turbidity level maintained < 10 NTU. 2)

130 Effect of turbidity (0 – 100 NTU): adjusted using SiO2 (0 – 0.00625%; w/v) with the COD level

131 maintained < 10 ppm. 3) Effect of COD (0, 1000 and 2000 ppm) and turbidity (0, 60 and 100

132 NTU): adjusted using tomato extract and SiO2. 4) Effect of the type of fresh produce extract

133 (blueberry, tomato and lettuce): COD and turbidity levels were adjusted using different produce

134 extracts and SiO2 to a condition similar to industrial operation (600 NTU and 2000 ppm COD).

135 The 10 L inoculated wash water with different levels of COD and turbidity was treated

136 by the overhead WUV configuration for up to 2 min, and it was sampled for Salmonella

137 population every 0.5 min. The intensity of UV was recorded at the surface of wash water and the

138 intensity was 23 – 28 mW/cm2. For each sample, 2 mL of water was taken out, added into 2 mL

139 of Dey-Engley (D/E) neutralizing broth and analyzed for Salmonella counts using standard plate

140 count method with tryptic soy agar supplemented with 0.6% of yeast extract, 100 μg/mL of

141 nalidixic acid and 100 μg/mL of streptomycin as described by Huang et al. (2018).

7
142 2.4. Comparisons of the efficacy of 3 WUV configurations on inactivation of Salmonella on

143 blueberry, tomato, and lettuce shreds

144 The fresh produce samples were dip-inoculated with the Salmonella cocktail following

145 the protocol described by Huang et al. (2018). Briefly, fresh produce samples were dipped in the

146 Salmonella cocktail for 2 min while being stirred. The inoculated samples were dried for 2 h at

147 room temperature, packed in plastic bags and kept at 4 °C overnight. The samples (3000 g of

148 blueberries, 3000 g of grape tomatoes, or 800 g of iceberg lettuce) were then placed in 10 L

149 turbid wash water (2000 ppm of COD and 100 NTU of turbidity). For tomatoes, the temperature

150 of wash water was adjusted to 30 °C to meet the FDA regulation (FDA, 2018) and for

151 blueberries and lettuce shreds, the temperature was adjusted to 4 °C following commercial

152 practice. For blueberries, the COD and turbidity were adjusted using blueberry extracts (2%) and

153 SiO2 (0.00375%). For tomatoes, the COD and turbidity were adjusted using tomato extracts (4%)

154 and SiO2 (0.00375%). For lettuce, the COD and turbidity were adjusted using lettuce extracts

155 (6%). The levels of COD and turbidity were determined as described by Huang, Sido, Huang, &

156 Chen (2015). The produce samples were stirred by circulating wash water while being treated by

157 the 3 WUV configurations for 2 min. In addition, a 2-min 10 ppm free chlorine wash (~ 10 ppm

158 free chlorine at the end of washing), which was commonly used in the industry for fresh produce

159 washing, was also tested for comparison purposes. The UV intensities were measured at the level

160 of wash water surface without any water loaded, and the distance between UV meter sensor and

161 the overhead UV lamps or the submersible UV lamp were ~20 cm and ~ 15 cm, respectively.

162 The intensities of overhead UV treatments and the submersible UV treatments were 23 – 28 and

163 ~ 4 mW/cm2, respectively. No apparent change of color or texture of fresh produce was found by

164 visual observation after treatments. After UV treatments, 2 mL of water was analyzed for

8
165 Salmonella counts as described in Section 2.3. To determine the Salmonella population in

166 produce samples, 100 g of blueberries, 100 g of grape tomatoes, and 15 g of iceberg lettuce were

167 randomly sampled, homogenized with D/E broth and quantified as described by Huang et al.

168 (2008).

169 2.5 Statistical analysis

170 The colony counts were converted to Log10 CFU/mL or Log10 CFU/g and presented in

171 mean ± standard deviation. Log reduction was calculated by subtracting the log count of treated

172 samples from untreated samples. Each treatment was repeated for at least 3 times independently.

173 JMP Pro (version 13, Cary, NC) was utilized to conduct the statistical analyses. Statistically

174 significant differences were identified based on a P-value of 0.05 via one-way analysis of

175 variance and multiple comparison was conducted using the Tukey’s HSD test.

176 3. Results

177 3.1 Influence of turbidity and COD on inactivation of Salmonella in simulated wash water

178 by UV

179 The inactivation of Salmonella in wash water by UV using the overhead configuration is

180 shown in Figure 3 (under the influence of COD) and Figure 4 (under the influence of turbidity).

181 Under all circumstances tested in this study, WUV treatments were able to completely eliminate

182 Salmonella inoculated in wash water (< 2 CFU/mL) within 2 min. The UV inactivation effect of

183 Salmonella in wash water showed a negative relationship with the increase of COD. With the 30

184 s and 60 s treatments, UV inactivated significantly higher (P < 0.05) amount of Salmonella in the

185 0 ppm group than in 1500, 2000, and 2500 ppm groups. As for the influence of turbidity, no

9
186 significant difference (P > 0.05) was found among all turbidity levels at all treatment times.

187 Additionally, WUV reduced Salmonella in wash water by ~ 5 log in 30 s at all turbidity levels.

188 The inactivation of Salmonella in simulated wash water with different combinations of

189 turbidity and COD by WUV is shown in Figure 5. The only significant difference was found at

190 60 s. The groups of 2000 ppm/60 NTU and 2000 ppm/100 NTU showed significantly lower (P <

191 0.05) Salmonella inactivation in wash water than the groups of 1000 ppm/60 NTU and 1000

192 ppm/100 NTU. Additionally, at 60 s, groups with the same COD level showed almost same

193 Salmonella inactivation regardless of turbidity levels. Furthermore, the effect of the type of fresh

194 produce extract was examined (Figure 6). At 30 and 60 s, tomato and lettuce groups showed

195 significantly lower (P > 0.05) Salmonella inactivation in wash water than the blueberry group.

196 3.2 Comparisons of the efficacy of 3 WUV configurations on inactivation of Salmonella on

197 blueberry, tomato, and lettuce shreds

198 The three WUV configurations achieved similar Salmonella reduction for blueberries

199 (1.8 – 2.0 log) and tomatoes (2.4 – 2.9 log) (Table 1). For lettuce, significantly higher (0.9 log

200 higher; P < 0.05) number of Salmonella was inactivated by the Overhead and Dual WUV

201 configurations than the Submersible WUV configuration. It should be noted that the UV

202 intensity of the Submersible system was much lower than the other two WUV configurations.

203 For blueberries and tomatoes, no significant difference (P > 0.05) in Salmonella reduction was

204 found between the chlorine wash and the Overhead and Dual WUV configurations. However, the

205 Overhead and Dual WUV configurations achieved significantly higher Salmonella reduction

206 than the chlorine wash for lettuce.

10
207 As for the residual Salmonella in wash water (Table 1), chlorine wash was able to

208 maintain Salmonella below < 2 CFU/mL except for one replicate in the lettuce group.

209 Salmonella survivors of 0.8-4.7 log CFU/mL were found in wash water after three WUV

210 treatments. No significant difference (P > 0.05) of residual Salmonella on blueberry or tomato

211 was found among the three WUV treatments. For lettuce group, the Submersible WUV treatment

212 (4.7 log) showed significantly higher (P < 0.05) residual Salmonella than the Overhead and Dual

213 WUV treatments (3.1 log).

214 4. Discussion

215 As shown in Figure 3, the UV inactivation of Salmonella in wash waster was impaired as

216 the COD level increased at 30 and 60 s for 30- and 60-s treatments. Less Salmonella was killed

217 by UV when the COD level of wash water increased. Selma et al. (2008) showed a similar result

218 that a 20-min UV treatment reduced total coliforms in onion wash water (953.1 ppm of COD) by

219 ~1.5 log while a 10-min UV treatment could reduce total coliforms in escarole wash water (48.6

220 ppm of COD) by ~ 2.0 log. Huang et al. (2015) showed that more Salmonella survivors were

221 found in raspberry wash water after water-assisted pulsed light treatments when the

222 concentration of raspberry extract in wash water is higher (1100 to 2007 ppm of COD). One

223 possible reason could be that the higher concentration of fresh produce extract contained higher

224 quantities of UV-absorbing compounds and led to diminished UV decontamination ability (Gil,

225 Selma, López-Gálvez, & Allende, 2009). Another explanation could be the organic matters

226 helped injured Salmonella cells recover from the UV treatments (Shafaei, Bolton, & El Din,

227 2016).

228 As shown in Figure 4, the turbidity did not show an impact on Salmonella survival in

229 wash water while being treated by UV, which was contradictory to previous studies (Gil et al.,

11
230 2009). Jones, Worobo, & Smart (2014) tested UV on inactivation of E. coli, Salmonella and

231 Listeria monocytogenes in unfiltered surface water and found that these pathogens were easier to

232 be killed by UV in PBS (0.1 NTU) than in pond water (19.6 NTU) or creek water (3.9 NTU).

233 Koutchma, Keller, Chirtel, & Parisi (2004) also showed that UV treatments achieved higher E.

234 coli inactivation in apple juice at a turbidity level of 2400 NTU than 1400 NTU. The difference

235 between our study and previous studies might be attributed to the high UV dose (~0.75-2.8 J/cm2)

236 adopted in our study. Ignat et al. (2015) demonstrated that a UV-treatment of 0.04 J/cm2 could

237 reduce > 5 log of E. coli, Salmonella, and Listeria monocytogenes in fresh-cut salad wash water.

238 Despite the increase of turbidity from 0 to 100 NTU in our study, the high dose of UV irradiation

239 enabled itself to penetrate through wash water and inactivate Salmonella effectively.

240 As for the effect of different combinations of COD and turbidity on UV inactivation of

241 Salmonella in wash water (Figure 5), the main factor influencing the UV decontamination ability

242 was the COD level. In the 60-s treatment, higher Salmonella survivors were found in the 2000

243 ppm COD group than the ones in the 1000 ppm COD group, but no difference was found

244 between groups at turbidity of 60 NTU and 100 NTU. The aforementioned results (Figure 3 and

245 4) also corroborated with this finding. Additionally, we did not observe any synergistic impact

246 between COD and turbidity on the UV decontamination ability. Despite the addition of SiO2 to

247 increase the turbidity, similar amount of Salmonella survivors were found in wash water with

248 tomato extract and SiO2 (Figure 5) and in wash water with only tomato extract (Figure 3).

249 As shown in Figure 6, we found that at similar levels of COD and turbidity, when treated

250 by UV, the survival of Salmonella in wash water followed this order: 2% blueberry extract +

251 SiO2 < 4% tomato extract + SiO2 < 6% lettuce extract. One probable explanation was that even

252 though the COD levels are similar, the content of UV-absorbing matters in wash water with 2%

12
253 blueberry extract + SiO2 was probably higher than wash water with 4% tomato extract + SiO2

254 and wash water with 6% lettuce extract. This also indicated that the COD level could not

255 properly reflect the level of UV-absorbing matters. When using UV for fresh produce wash

256 water decontamination, the UV-absorbance of wash water should be included as a parameter for

257 wash water quality test.

258 Submersible UV lamps have been used for waste water decontamination (Trojan

259 Technologies, n.d.; United States Environmental Protection Agency, 1999). In this study, we

260 installed a submersible UV lamp to enhance the fresh produce decontamination ability.

261 Compared to the Overhead and the Dual WUV configurations, the Submersible WUV

262 configuration achieved similar Salmonella reductions on blueberry and tomato, but worse

263 reduction on lettuce (Table 1). However, the intensity of the submersible UV lamp (4 mW/cm2)

264 was much lower than the overhead WUV lamp (~25 mW/cm2), which indicated its potential in

265 decontamination for blueberry and tomatoes. As for lettuce, a possible reason might be that the

266 lettuce shreds had a tendency to get entangled onto the submersible UV lamp and block the UV

267 light. However, when we combined the Overhead WUV and the submersible WUV together to

268 produce the Dual WUV configuration, we did not observe significant (P > 0.05) higher

269 Salmonella reduction than both separate WUV configurations. This indicated that there was

270 some place shielding the Salmonella from both UV lamps. It also should be noted that, during

271 the washing with stirring pads, tomatoes could hit onto the submersible UV lamp and cause lamp

272 damage (all results presented in the study were conducted in the undamaged condition). Certain

273 protection methods of the submersible UV lamp, such as a mesh net, should be employed in the

274 future application.

13
275 In conclusion, UV is a promising intervention in decontamination of fresh produce wash

276 water. The Overhead WUV configuration was able to eliminate > 6 log of Salmonella suspended

277 in wash water containing very high levels of COD and turbidity within 2 min. Due to the cost of

278 water usage, it is a common practice for the fresh produce industry to reuse or to recirculate wash

279 water. Our results demonstrated that the Overhead WUV configuration could be used to

280 decontaminate spent wash water before they are reused for washing a new batch of fresh produce.

281 Less amount of Salmonella in wash water was killed by UV as the COD level increased from 0

282 to 2500 ppm, but not when the turbidity increased from 0 to 100 NTU. No synergistic or

283 antagonist effect between COD and turbidity on impacting UV inactivation of Salmonella in

284 wash water was found. However, the type of produce extract used to simulate actual wash water

285 condition had an impact on the Salmonella inactivation. The three WUV configurations showed

286 similar Salmonella reduction on blueberries (1.8-2.0 log) and tomatoes (2.4-2.9 log). However,

287 the Overhead and Dual WUV configurations achieved significantly higher (P < 0.05) Salmonella

288 reduction on lettuce than the Submersible WUV configuration.

289 Acknowledgements

290 This project was supported by the Agriculture and Food Research Initiative Competitive

291 Grants Program of the USDA National Institute of Food and Agriculture, NIFA Award No.:

292 2015-69003-23410.

293 References
294 Bennett, S. D., Sodha, S. V, Ayers, T. L., Lynch, M. F., Gould, L. H., & Tauxe, R. V. (2018).

295 Produce-associated foodborne disease outbreaks, USA, 1998–2013. Epidemiology and

296 Infection, 146(11), 1397–1406. https://doi.org/DOI: 10.1017/S0950268818001620

14
297 CDC. (2018). Multistate Outbreak of E. coli O157:H7 Infections Linked to Romaine Lettuce

298 (Final Update). Retrieved December 1, 2018, from https://www.cdc.gov/ecoli/2018/o157h7-

299 04-18/index.html

300 Centers for Disease Control and Prevention. (2018). Outbreak of E. coli Infections Linked to

301 Romaine Lettuce. Retrieved December 31, 2018, from

302 https://www.cdc.gov/ecoli/2018/o157h7-11-18/index.html

303 Coroneo, V., Carraro, V., Marras, B., Marrucci, A., Succa, S., Meloni, B., … Schintu, M. (2017).

304 Presence of Trihalomethanes in ready-to-eat vegetables disinfected with chlorine. Food

305 Additives & Contaminants: Part A, 34(12), 2111–2117.

306 https://doi.org/10.1080/19440049.2017.1382723

307 Cossu, A., Ercan, D., Tikekar, R. V., & Nitin, N. (2016). Antimicrobial Effect of Photosensitized

308 Rose Bengal on Bacteria and Viruses in Model Wash Water. Food and Bioprocess

309 Technology, 9(3), 441–451. https://doi.org/10.1007/s11947-015-1631-8

310 Fan, X., Huang, R., & Chen, H. (2017). Application of ultraviolet C technology for surface

311 decontamination of fresh produce. Trends in Food Science and Technology, 70, 9–19.

312 https://doi.org/10.1016/j.tifs.2017.10.004

313 FDA. (2018). Program Information Manual: Retail Food Protection Storage and Handling of

314 Tomatoes. Retrieved August 25, 2019, from

315 https://www.fda.gov/Food/GuidanceRegulation/RetailFoodProtection/IndustryandRegulator

316 yAssistanceandTrainingResources/ucm113843.htm

317 Food and Agriculture Organization of the United Nations. (2018). FAOSTAT. Retrieved

318 December 19, 2018, from http://www.fao.org/faostat

15
319 Garg, R., Abela, D., Tiwari, B., & Valdramidis, V. (2016). Potential industrial applications of

320 decontamination technologies for fresh produce. Food Hygiene and Toxicology in Ready-to-

321 Eat Foods, 313–336. https://doi.org/10.1016/B978-0-12-801916-0.00018-2

322 Gil, M. I., Selma, M. V., López-Gálvez, F., & Allende, A. (2009). Fresh-cut product sanitation

323 and wash water disinfection: Problems and solutions. International Journal of Food

324 Microbiology, 134(1–2), 37–45. https://doi.org/10.1016/j.ijfoodmicro.2009.05.021

325 Goodburn, C., & Wallace, C. A. The microbiological efficacy of decontamination methodologies

326 for fresh produce: A review, 32 Food Control § (2013). Elsevier.

327 https://doi.org/10.1016/j.foodcont.2012.12.012

328 Guo, S., Huang, R., & Chen, H. (2017). Application of water-assisted ultraviolet light in

329 combination of chlorine and hydrogen peroxide to inactivate Salmonella on fresh produce.

330 International Journal of Food Microbiology, 257, 101–109.

331 https://doi.org/10.1016/j.ijfoodmicro.2017.06.017

332 Haute, Sam van, Sampers, I., Jacxsens, L., & Uyttendaele, M. (2015). Selection Criteria for

333 Water Disinfection Techniques in Agricultural Practices. Critical Reviews in Food Science

334 and Nutrition, 55(11), 1529–1551. https://doi.org/10.1080/10408398.2012.705360

335 Hoffmann, S. A., Maculloch, B., & Batz, M. (2015). Economic burden of major foodborne

336 illnesses acquired in the United States. United States Department of Agriculture, Economic

337 Research Service.

338 Huang, R., de Vries, D., & Chen, H. (2018). Strategies to enhance fresh produce

339 decontamination using combined treatments of ultraviolet, washing and disinfectants.

340 International Journal of Food Microbiology, 283, 37–44.

16
341 https://doi.org/10.1016/J.IJFOODMICRO.2018.06.014

342 Huang, Y., Sido, R., Huang, R., & Chen, H. (2015). Application of water-assisted pulsed light

343 treatment to decontaminate raspberries and blueberries from Salmonella. International

344 Journal of Food Microbiology, 208, 43–50.

345 https://doi.org/10.1016/j.ijfoodmicro.2015.05.016

346 Ignat, A., Manzocco, L., Bartolomeoli, I., Maifreni, M., & Nicoli, M. C. (2015). Minimization of

347 water consumption in fresh-cut salad washing by UV-C light. Food Control, 50, 491–496.

348 Retrieved from http://linkinghub.elsevier.com/retrieve/pii/S095671351400557X

349 Jones, L. A., Worobo, R. W., & Smart, C. D. (2014). UV Light Inactivation of Human and Plant

350 Pathogens in Unfiltered Surface Irrigation Water. https://doi.org/10.1128/AEM.02964-13

351 Koutchma, T. Advances in Ultraviolet Light Technology for Non-thermal Processing of Liquid

352 Foods, 2 Food and Bioprocess Technology § (2009). Springer-Verlag.

353 https://doi.org/10.1007/s11947-008-0178-3

354 Koutchma, T., Keller, S., Chirtel, S., & Parisi, B. (2004). Ultraviolet disinfection of juice

355 products in laminar and turbulent flow reactors. Innovative Food Science & Emerging

356 Technologies, 5(2), 179–189. https://doi.org/10.1016/J.IFSET.2004.01.004

357 Luo, Y., Nou, X., Millner, P., Zhou, B., Shen, C., Yang, Y., … Shelton, D. (2012). A pilot plant

358 scale evaluation of a new process aid for enhancing chlorine efficacy against pathogen

359 survival and cross-contamination during produce wash. International Journal of Food

360 Microbiology, 158(2), 133–139. https://doi.org/10.1016/j.ijfoodmicro.2012.07.008

361 Nou, X., Luo, Y., Hollar, L., Yang, Y., Feng, H., Millner, P., & Shelton, D. (2011). Chlorine

17
362 Stabilizer T-128 Enhances Efficacy of Chlorine against Cross-Contamination by E. coli

363 O157:H7 and Salmonella in Fresh-Cut Lettuce Processing. Journal of Food Science, 76(3),

364 M218–M224. https://doi.org/10.1111/j.1750-3841.2011.02046.x

365 Rekhy, R., & McConchie, R. (2014). Promoting consumption of fruit and vegetables for better

366 health. Have campaigns delivered on the goals? Appetite, 79, 113–123.

367 https://doi.org/10.1016/J.APPET.2014.04.012

368 Sayed, S. K. I., El-Ezaby, K. H., & Groendijk, L. (2005). Treatment of potato processing

369 wastewater using a membrane bioreactor. In Ninth International Water Technology

370 Conference IWTC9 (pp. 53–68).

371 Selma, M. V., Allende, A., López-Gálvez, F., Conesa, M. A., & Gil, M. I. (2008). Disinfection

372 potential of ozone, ultraviolet-C and their combination in wash water for the fresh-cut

373 vegetable industry. Food Microbiology, 25(6), 809–814.

374 https://doi.org/10.1016/j.fm.2008.04.005

375 Shafaei, S., Bolton, J. R., & El Din, M. G. (2016). Effect of Environmental Factors on

376 Photoreactivation of Microorganisms under Indoor Conditions. World Academy of Science,

377 Engineering and Technology, International Journal of Environmental, Chemical,

378 Ecological, Geological and Geophysical Engineering, 10(5), 560–563.

379 Trojan Technologies. (n.d.). UV Disinfection for Wastewater.

380 Tudela, J. A., López-Gálvez, F., Allende, A., Hernández, N., Andújar, S., Marín, A., … Gil, M. I.

381 (2019). Operational limits of sodium hypochlorite for different fresh produce wash water

382 based on microbial inactivation and disinfection by-products (DBPs). Food Control, 104,

383 300–307. https://doi.org/10.1016/j.foodcont.2019.05.005

18
384 United States Environmental Protection Agency. (1999). Wastewater Technology Fact Sheet

385 Ultraviolet Disinfection. Retrieved December 31, 2018, from

386 https://www3.epa.gov/npdes/pubs/uv.pdf

387 Van Haute, S., Luo, Y., Sampers, I., Mei, L., Teng, Z., Zhou, B., … Millner, P. (2018). Can UV

388 absorbance rapidly estimate the chlorine demand in wash water during fresh-cut produce

389 washing processes? Postharvest Biology and Technology, 142, 19–27.

390 https://doi.org/10.1016/J.POSTHARVBIO.2018.02.002

391 Van Haute, S., Tryland, I., Veys, A., & Sampers, I. (2014). Wash water disinfection of a full-

392 scale leafy vegetables washing process with hydrogen peroxide and the use of a commercial

393 metal ion mixture to improve disinfection efficiency. Food Control, 50, 173–183.

394 https://doi.org/10.1016/j.foodcont.2014.08.028

395 Van Haute, Sam, Sampers, I., Holvoet, K., & Uyttendaele, M. (2013). Physicochemical quality

396 and chemical safety of chlorine as a reconditioning agent and wash water disinfectant for

397 fresh-cut lettuce washing. Applied and Environmental Microbiology, 79(9), 2850–2861.

398 https://doi.org/10.1128/AEM.03283-12

399

400

19
401 Figures
402

403

404 Figure 1. Overhead WUV configuration setup. (A) A schematic diagram. (B) A picture showing

405 tomatoes being washed.

406

20
407

408 Figure 2. Dual and Submersible WUV configurations. (A) A schematic diagram of the Dual

409 WUV configuration. (B) A schematic diagram of the Submersible WUV configuration. (C) A

410 picture of the Submersible WUV configuration.

411

21
412

413 Figure 3. Inactivation of Salmonella in wash water using the Overhead WUV configuration –

414 Effect of COD levels (0 – 2500 ppm). Initial concentration of Salmonella was ~ 2 × 106 CFU/mL

415 and the limit of detection was 2 CFU/mL. Error bars represent one standard deviation. Bars at the

416 same treatment time with same uppercase letter are not significantly different (P > 0.05).

417

22
418

419 Figure 4. Inactivation of Salmonella in wash water using the Overhead WUV configuration –

420 Effect of turbidity levels (0 – 100 NTU). Initial concentration of Salmonella was ~ 2 × 106

421 CFU/mL and the limit of detection was 2 CFU/mL. Error bars represent one standard deviation.

422 Bars at the same treatment time with same uppercase letter are not significantly different (P >

423 0.05).

424

23
425

426 Figure 5. Inactivation of Salmonella in wash water using the Overhead WUV configuration –

427 Effect of COD (1000 and 2000 ppm) and turbidity (60 and 100 NTU) levels. Initial concentration

428 of Salmonella was ~ 2 × 106 CFU/mL and the limit of detection was 2 CFU/mL. Error bars

429 represent one standard deviation. Bars at the same treatment time with same uppercase letter are

430 not significantly different (P > 0.05).

431

24
432

433 Figure 6. Inactivation of Salmonella in wash water using the Overhead WUV configuration –

434 Effect of type of fresh produce extract. The COD and turbidity levels were maintained at ~ 2000

435 ppm and 60 NTU, respectively. Three types of fresh produce were tested: 1) 2% blueberry

436 extract/0.00375% SiO2, 2) 4% tomato extract/0.00375% SiO2 and 3) 6% lettuce extract. Initial

437 concentration of Salmonella was ~ 2 × 106 CFU/mL and the limit of detection was 2 CFU/mL.

438 Error bars represent one standard deviation. Bars at the same treatment time with same uppercase

439 letter are not significantly different (P > 0.05).

440

25
441 Tables

442 Table 1. Comparisons of the efficacy of 3 WUV configurations on inactivation of Salmonella on

443 blueberry, tomato, and lettuce shreds. Dip-inoculated fresh produce was washed in turbid tap

444 water (2000 ppm of COD and 100 NTU of turbidity) while being treated by the Overhead WUV,

445 Submersible WUV or Dual WUV configurations for 2 min. Data represent mean of three

446 replicates ± standard deviation.

Treatments Blueberry Tomato Lettuce

Log reduction of Salmonella on fresh produce (log CFU/g)
Chlorine wash 2.2 ± 0.1 Bb 3.3 ± 0.2 Ac 0.8 ± 0.3 Aa
Overhead WUV 1.8 ± 0.1 Ab 2.5 ± 0.2 Ac 1.5 ± 0.1 Ba
Submersible WUV 1.9 ± 0.1 ABb 2.4 ± 0.4 Ab 0.6 ± 0.2 Aa
Dual WUV 2.0 ± 0.1 ABa 2.9 ± 0.4 Ab 1.5 ± 0.1 Ba
Residual Salmonella in wash water (log CFU/mL)
Chlorine wash <0.3 ± 0.0 (3/3) Aa <0.3 ± 0.0 (3/3) Aa <0.3 ± 0.0 (2/3) Aa
Overhead WUV 2.0 ± 0.4 Bab <0.8 ± 0.9 (1/3) Aa 3.1 ± 0.1 Bb
Submersible WUV 2.4 ± 0.2 Ba 1.5 ± 0.6 Aa 4.7 ± 0.3 Cb
Dual WUV 2.2 ± 0.3 Bb 0.7 ± 0.5 Aa 3.1 ± 0.1 Bc

447 Within the same category of “Log reduction of Salmonella on fresh produce (log CFU/g)” or

448 “Residual Salmonella in wash water (log CFU/mL)”, data in the same column followed by the
449 same uppercase letter are not significantly different (P > 0.05).

450 Data in the same row followed by the same lowercase letter are not significantly different (P >

451 0.05).
452 The initial inoculation levels of Salmonella on blueberry, tomato, and lettuce were 6.1 ± 0.0, 4.1

453 ± 0.0, and 6.6 ± 0.1 log CFU/g, respectively.

454 The detection limits of Salmonella on blueberry, tomato, and lettuce were 3, 3, and 11 CFU/g,
455 respectively.

456 The fraction numbers in parentheses represent the number of replicates below the detection limit

457 out of the number of replicates. For replicates below the detection limit, the detection limit was
458 used in the calculation of means and standard deviations.

26