Food Protection Trends, Vol. 27, No. 9, Pages 678-683Copyright 2007, International Association for Food Protection6200 Aurora Ave., Suite 200W, Des Moines, IA 50322-2864 Q

InIina!ioflI ASuCIaIIOfl lvi Food Protection.

Evaluation of CleaningTreatments for Almond-contact Surfaces in Hullingand Shelling FacilitiesWEN-XIAN DU,' MICHELLE D. DANYLUK, 2 and LINDAJ. HARRIS3'USDA-ARS-WRRC, Processed Foods Research Unit, 800 Buchanan St.,Albany, CA 94710, USA; 2University of Florida,Dept. of Food Science and Human Nutrition, Citrus Research and Education Center, 700 Experiment Station Road,Lake Alfred, FL 33850, USA, and 3 Dept. of Food Science and Technology, University of California-Davis, One Shields Ave.,Davis, CA 95616, USA

SUMMARYAfter harvest, almond hulls and shells are separated from kernels in specialized huller-sheller

(HS) facilities. This study evaluated various cleaning and sanitizing treatments on contact surfacestypically found in HS facilities. Dust (hull, shell, and soil particulates) was collected from an HS facilityand applied to samples of new and worn conveyor belting and painted and unpainted galvanized steel.Dust-contaminated surfaces (100 cm') were swabbed before and after blowing with air for 30 s and/or wetting with water, an aqueous commercial cleaner, or isopropyl alcohol quaternary ammonium(IPAQUAT) sanitizer. Aerobic plate counts (APCs) and, in some cases, adenosine triphosphate (ATP)levels were determined. Combinations of air, commercial cleaner, and IPAQUAT significantly reducedAPCs and ATP levels; however, the correlation between APCs and ATP levels was poor. The use ofwater or aqueous cleaners is not recommended for HS facilities unless complete dust removal canbe assured and there is adequate time for thorough drying of equipment (e.g., post- or pre-season).Air blowing reduced APCs by 10 fold (on worn and new belting) to 100 fold (on unpainted andpainted steel). In the laboratory, applying an IPAQUAT sanitizer after air blowing reduced APCs byan additional 10 fold on belting or 100 fold on steel surfaces. However, when this same treatmentwas evaluated in a commercial HS facility, the APCs were not significantly reduced on any of thesurfaces tested.

A peer-reviewed article

*Author for correspondence: 530.754.9485; Fax: 530.752.4759E-mail: tlharris@ucdavis.edu

678 FOOD PROTECTION TRENDS I SEPTEMBER 2007

INTRODUCTION

Huller and huller-sheller (HS) 6icili-ties receive harvested almonds and removethe hull only or the hull and shell to pro-duce in-shell or shelled almond kernels,respectively. Large volumes of "dust"-composed primarily of shell and/or hullparticulates and soil collected duringharvest—are generated by the hulling andshelling operations (10. This dust, whichis ubiquitous in huller and HS facilities,coats conveyor and equipment surfacesshortly after almonds enter the facility andthe hulling and shelling begin.

Almond-contact surfaces (e.g., pro-cessing equipment and conveyors) in HSfacilities constitute important points itwhich microbial contamination couldspread within and between productionlots. HS facilities are in operation forabout 5 months during the almondharvest, which generally runs from Julythrough November. These facilities havenot traditionally employed in-seasoncleaning and sanitizing programs beyondgeneral housekeeping. Some of the HSdust is ducted to a cyclone or fabric filterfor collection and disposal. Dust notcaptured by the cyclones is removed byperiodic air blowing of the ceilings, walls,equipment, and floors, and then collectingand disposing of the dust (1).

Many HS facilities are not con-structed to meet the 21 Code of FederalRegulations (CFR) 110 Current GoodManufacture Practices (GM Ps) becausethey have not traditionally been classifiedas food processors (2). In addition, theequipment is not designed or intendedto be broken down for regular cleaningexcept at the end of the season or whilein repair.

After the 2000 to 2001 outbreak ofsalmonellosis associated with raw almondconsumption (5), HS facilities were askedto register as food processing facilities andto substantially increase the amount andfrequency of in-season cleaning and sani-tation, and GMPs for cleaning and sanitiz-ing HS facilities were then established (1).However, data validating the efficacy ofsanitation in dry food facilities is difficultto find. Most textbooks on sanitationdevote only minor chapters to dry facilitysanitation methods, primarily for cereals,grains, and their products (8, 9).

The overall objective of this researchwas to provide the almond HS industrywith data on the efficacy of cleaningand sanitizing methods specific to theirequipment and facilities. The followingstudies were carried out: (i) evaluatingvarious cleaning and sanitizing treat-ments on surfaces typically found in HSfacilities, for potential use in the off- orpre-season; (ii) evaluating the in-seasonindustry practice of removing dust withair blowing followed by application ofan alcohol-based quaternary ammoniumsanitizer for reducing the microbial loadson kernel-contact surfaces; and (iii) evalti-ating the efficacy of this sanitizing methodri an almond HS ftcilitv.

MATERIALS AND METHODS

Almond HS dustAs previously reported by our labo-

ratory (3), the aerobic plate counts(APCs) of almond dusts (hull, shell, andsoil particulate material) collected fromvarious locations in two HS facilitiesranged from 5.8 to 6.8 log CFU/g. Forthis study, a single dust sample (with anAPC of 6.8 ± 0.1 log CFU/g) collectedfrom one of these facilities was selectedas the representative dust to use for thecontamination of surfaces. The dustsample was stored at ambient temperature(23 ± 3°C) in a sealed polyethylene bag(30.5 x 30.5 cm, Bitran; Com-Pac Int.,Carbondale, II,), which was placed insidea scaled plastic tub.

SurfacesSamples of conveyor belting

monly used in the HS industry, includingunpainted new galvanized steel, and newor worn smooth belting or worn scallopedbelting, were obtained from HS facilities.Each surface type was cut into 15 x 15cm squares (225 cm 2). Some of the newgalvanized steel squares were painted foruse in the study.

A standard washing procedure wasdeveloped to prepare the surface samplesbefore treatment (dust contamination andcleaning/sanitizing). Squares were rinsedwith running distilled water for 1 mm,sanitized with 75% ethanol by spraying tocompletely wet the surface, blotted withfacial tissues, and then air dried at roomtemperature. Surfaces washed by this pro-cedure were free of visible dust and hadAPCs of< 2.4 log CFU/lOO cm2.

A contamination protocolfor surfaces

In initial experiments, which evalu-ated more traditional wet-processingcleaning and sanitizing protocols, 5 mg ofHS dust was applied to surface samples.For subsequent experiments, 1 g of dustwas applied; this quantity more closelyrepresented dust levels observed in HSfacilities during operation. A stainlesssteel fine-mesh strainer was used to evenlydistribute 5 mg or 1 g of dust onto eachstirface as it was lying horizontally. Allthe dust-contaminated surfaces were leftundisturbed for 10 s before applying atreatment; air blowing, tap water, cleaner,sanitizer, or various combinations ofthese treatments were used, as describedbelow.

The movement and weight of al-monds on conveyor belts and equipmentsurfaces may compress the dust, makingits subsequent removal more difficult. Todetermine the effect of compacted dust onsurface cleaning, I g of dust was spreadover various surface squares and the samesurface type was placed face down ontop of the dust. A 1-I beaker filled withwater to a weight of 1 kg was placed onthe back side of the top square and leftfor 18 to 24 h to press the dust onto thetest surface. The top square was removedand a cleaning or sanitizing treatment wasapplied to the bottom surface square, asdescribed below.

Cleaner and sanitizer

A commercial cleaner (HC- 10Chlorinated Kleer-Mor; Ecolab, St. Paul,MN) that is representative of those usedin wet food-processing facilities, and analcohol-based quaternary ammonium(IPAQUAT) sanitizer (Alper D2 [58.6%isopropyl alcohol]; Best Sanitizers Inc.,Penn Valley, CA) used in HS facilities atthe time of the study, were evaluated. Thecleaner was prepared by adding 1 package(230 g) to 15 liter of warm water (49°C)as instructed on the label. The sanitizerwas supplied at full strength (200 ppm)and was used at this concentration.

Cleaning and sanitizingtreatment protocols

For the air blowing treatment, eachdust-contaminated surface was placed

PF

SEPTEMBER 2007 1 FOOD PROTECTION TRENDS 679

Aerobic plate count (log CFU/cm2)'

Surface without dust Surface with dust 30 s air + 60 sSurface Thickness (cm)(negative control)(positive control)30 s airIPAQUATNew steel"

0.151.1 ±0.3A

5.1 ±0.4C

3.4±0.2B1.5±0.5A

Painted steel'

0.151.6±0.IA

5.5 ± 0.8 C

3.5±0.2B1.2±0.9A

4.1±0.3B2.9±0.2B

New conveyor belting

Smooth, yellow"'

Worn conveyor belting

Smooth, white"

Smooth, red, thick1

Smooth, red, thin1

Scalloped, blackt

Scalloped, white

0.5

0.4

0.6

0.2

0.3-0.5

0.3-0.7

1.6 ± 0.4A

.5 ± 0.2A

1.4 ± 0.6A

1.7 ± 0.4A

2.4 ± 0.5A

2.2 ± 0.1 A

5.0 ± 0.0 D

5.3 ± 0.7 D

5.3 ± 0.4 D

5.5 ± 0.0 D

5.3 ± 0.4 D

5.4 ± 0.4 D

3.9 ± 0.0 C

3.8 ± 0.4 C

4.1 ±0.1 C

4.0 ± 0.2 C

4.1 ±0.1 C

3.2 ± 0.0 B

2.8 ±0.1 B

3.1 ±0.0 B

3.2 ±0.1 B

3.2 ± 0.0 B

'Means (n = 6) with different letters in the same row are significantly different (P < 0.05).

'Manufacturer not available; 'Goodyear, Marysville, OH; dThermoid Bellefontaine, OH;

'Scandura Scottsdale, GA; Nobet (location not available); tBelt Concepts, Marysville, OH.

horizontally in the bottom of a clear poly-propylene bag (61 x 91 cm; Fisher, FairLawn, NJ) inside a large plastic tub. TygonR-3603 tubing (4.8min I.D. and 14 minO.D.; Fisher Scientific) was used to blowin-house air across the surface for 30 s bymanually moving the tubing from side toside with one hand, while sealing the topof the bag with the other hand to preventthe dust from flying out of the bag.

For the liquid cleaning treatments,each dust-contaminated surface wasplaced vertically (to prevent pooling ofliquid) into a plastic dishwashing basketset in a sink. Tap water or cleaner (200ml) was slowly poured from the topover the entire surface. Alternatively, theIPAQUAT was sprayed over the surface,using five consecutive sprays to ensure thatthe surface was wetted for a minimum ofI mm, as per the manufacturer's instruc-tions. All the liquid-treated surfaces wereheld at ambient temperature to dry (ap-proximately 10 mm) before swabbing.

Various combinations of the cleaningtreatments were also examined for theirefficacy in reducing microbial loads on thesurfaces. These combination treatments

were carried out following the protocolsabove, one following the other. For ex-ample, the air + water + cleaner treatmentconsisted of air blowing, followed directlyby the tap water treatment and then fol-lowed by the cleaner treatment.

Swab sampling to evaluatesurface microbial load

A 100-cm 2 area of each dust-con-taminated surface was swabbed usinga standard technique (4, 6, 7) beforeand after various cleaning treatments orcombinations of treatments. A disposablesterile template (Weber Scientific, Hamil-ton, NJ) with a 10 x 10 cm opening wasplaced over each surface to ensure thata consistent area was swabbed. Com-mercial swabs in 7 ml neutralizing solu-tion (buffered peptone water containinglecithin and Tween 80; Weber Scientific,Hamilton, NJ) were used. Each swabsample was vortexed at high speed for5 s and the buffer was serially diluted(10 fold) in Butterfield's phosphate buffer(BPB). Appropriate dilutions were platedonto Petrifilm aerobic count (AC) plates

(3M Microbiology Products, St. Paul,MN). Swabs of clean surfaces (no dust)were used as negative (uninoculated) con-trols, and swabs of surfaces covered withdust (no cleaning treatment) were used aspositive (inoculated) controls.

ATP-bioluminescence toevaluate surface cleanliness

A commercial ATP-bioluminescencemonitoring system (LUMinatorT; CharmSciences, Lawrence, MA) was used ondust-contaminated unpainted steel sur-faces before and after cleaning. Standardswabbing techniques for direct surfacesampling were carried out using the Pock-etSwab (Charm Sciences). Results wereexpressed in relative light units (RLU).

Swab sampling of almond HSfacility

Commercial swabs (Weber Scien-tific) were used to evaluate the efficacy ofa representative sanitation routine usedin one almond HS facility. To cleanequipment surfaces, this facility removeddust with pressurized air and then ap-

680 FOOD PROTECTION TRENDS I SEPTEMBER 2007

12 3 45 6 7 8 9 10 110

6

5('1E0400

U-C)C)o2

FIGURE I. Cleaning treatment efficacy as measured by aerobic plate count (U)or ATP bioluminescence (1_i) on unpainted steel surfaces contaminated with 5 mg ofdust. Treatments included: (I) uriinoculated control, (2) inoculated control, (3) 30 s air,(4) 200 ml water, (5) 200 ml cleaner, (6) 60 S IPAQUAT, (7) 30 s air + 200 ml water, (8)30 s air + 200 ml cleaner, (9) 30 s air + 200 ml water + 200 ml cleaner, (10)30 s air +200 ml cleaner + 60 s IPAQUAT, (II) 30 s air + 200 ml water + 200 ml cleaner + 60 sI PAQUAT.

Cleaning Treatment

6

50(0

I-C

) -'Jo0

2

1

0

plied an IIAQUAT sanitizer, as recom-mended in the almond Huller/ShellerGood Manufacturing Practices andSanitation Manual (1). Duplicate swabsamples were collected from each of 20sample sites before and after the facility'ssanitation routine. Swabbing consistedof 10 horizontal strokes and 10 verticalstrokes within an approximate 10 x 10cm sampling area of various surfaces, in-eluding belting and steel (unpainted andpainted) equipment. The swab sampleswere immediately placed in a cooler withfreezer packs and were transported to thelaboratory for microbiological analysis.All swab samples were analyzed within24 h of sampling. Each swab samplewas vortexed vigorously, and serially di-luted (10 fold) in BPB, after which 1 mlwas plated onto single Petrifllm AC andE. coli/coliform plates. The facility wasswabbed twice during one harvest seasonand the data from each surface type werecombined.

Statistical analysis

Data were analyzed using StatisticalAnalysis System (version 8.2) software(SAS Institute, Cary, NC). Analysis ofvariance by the General Linear Model(GLM) procedure and Duncan's multiplerange tests were used to determine thesignificant differences (P < 0.05) amongtests or methods.

RESULTS AND DISCUSSIONIt is common for HS facilities to

operate continuously during the harvestseason, generating large volumes of dustthat cover all surfaces soon after start up.Under conditions of high dust volumeand low moisture, it is recommendedthat processing equipment not be cleanedusing aqueous products (9). Introducingmoisture to this environment may increasethe risk for almond contamination, byproviding niches for microorganisms,such as Salmonella, to grow (3). Duringthe off-season, however, when equipmentcan be disassembled and completely dried,

the use of an aqueous cleaner, in combina-tion with other cleaning treatments, maybe beneficial.

Cleaning treatment efficacy onunpainted steel surfaces, as evaluated byAPC and ATP, is shown in Fig. 1. Afterapplication of 5 mg of dust, the APC was4.2 ± 0.2 CFU/100 cm' (Fig. 1-bar 2), andfollowing air blowing, the APC droppedby 1.9 log CFU/100 cm 2 , to 2.3 ± 0.1 logCFU/100 cm' (Fig. 1-bar 3).

Significant (P < 0.05) reductionsafter air blowing, to the level of theuninoculated control, were observed forall treatment combinations, with theexception of 30 s air + 200 ml water (Fig.1-bar 7). The greatest reductions in APC(3.6 log CFU/ 100 cm 2) were observed fol-lowing the combination of 30 s air + 200ml cleaner + 60 s IPAQUAT (Fig. 1-bar10). Applying a force to the surfaces tocompact the dust did not have a signifi-cant effect (P> 0.05) on APCs, before orafter cleaning or sanitizing, compared tosurfaces where no force was applied (datanot shown).

The ATP readings did not correlatewell with the APCs (Fig. 1), particularlyfor treatment with 60 s IPAQUAI (Fig1-bar 6; APC low/ATP high) and 30air + 200 ml water (Fig. 1-bar 7; APChigher/ATP equivalent to uninoculatedcontrol). For this reason, the ATP-biolu-minescence system was not used in furtherexperiments.

Under laboratory conditions, thenew and worn surfaces contaminatedwith 1 g of dust had APCs of 5.0 to 5.5log CFU/100 cm before sanitizing (Table1). Air blowing for 30 s reduced APCs by1 to 2 log CFU/100 cm2 , depending onthe surface, with the greatest reductionsobserved on steel surfaces. Applying analcohol-based sanitizer (IPAQUAT) afterair blowing reduced APCs further, by 1log CFU/100 cm- on belting surfaces andby 2 log CFU/100 cm 2 on steel surfaces.Air blowing followed by IPAQUAT wasan equally effective treatment on all wornbelting surfaces.

At the commercial HS facility tested,samples collected from equipment sur-faces before air blowing and applicationof IPAQUAT had average APCs of 4.2to 4.8 log CFU/100 cm 2 , similar to theinoculated controls prepared for use inthe laboratory (Fig. 2). Coliform countswere 1.8 ± 0.9, 2.5 ± 0.7, and 1.3 ± 0.5log CFU/100 cm 2 for steel, painted steel,

SEPTEMBER 2007 1 FOOD PROTECTION TRENDS 681

FIGURE 2. Microbial populations on surfaces at a commercial almond HS facilitybefore •) and after (1) pressurized air blowing followed by application of an alcohol-based sanitizer (IPAQUAT). Limit of detection (0.8 log CFU/ 100 cm2).

APC

LL

Steel (n=8)

5U0a

U.

2

11filir'LlSteel (n=8)

7

Coliform6

S

Ua4a

U-3

Steel (n=8)

0

Painted Steel (n =4)Smooth Black Belting

(n=28)

• Before SanitationC3 After Sanitation

fT

Painted Steel (n =4)Smooth Black Belting

(n=28)

• Before Sanitationo After Sanitation

iIiPainted Steel (n =4)Smooth Black Belting

(n=28)

and smooth black belting, respectively.The presumptive E. coti counts were:S: 2.0 log CFU/100 cm2 in all samplescollected. No significant changes in APCs,coliform counts, or F. co/i counts wereseen following sanitation of the HS facilityfor any of the three surfaces tested (steel,painted steel, or smooth black belting).Under commercial conditions, air wasblown onto equipment starting at thehighest point and working down, so thedurations during which air was appliedto individual surfaces varied but weregenerally less than 30 s. In the labora-tory, air was applied to relatively smallsample surfaces for 30 s. The difficulty inuniformly removing dust and applying1PAQUA1' under commercial conditionsmax', in part, provide an explanation forthe observed lack of microbial reductioncompared with results of the labora-tory experiments. In addition, dust levelsequivalent to those present prior to clean-ing and sanitizing were observed withinminutes of resuming production.

RECOMMENDATIONS

Iraditional wet facilit y cleaningand sanitizing methods using water andchemical cleaners were very effective in re-ducing APCs tinder laboratory conditionsand may be appropriate for HS facilitiesin the off- or pre-season, when all partsof the equipment are accessible, dust canbe completely removed, and time neededfor thorough drying is available. This typeof cleaning is not recommended duringthe processing season when water mayprovide harborage for pathogens in inac-cessible areas of equipment. Based on theinconsistent results observed in this study,ATP is not recommended for monitoringsanitation in HS facilities.

Current recommended in-season HScleaning and sanitation practices (1) sig-nificantly reduced microbial populationson both new and worn kernel-contactsurfaces under laboratory conditions,but not when evaluated in a commercialfacility. The volume of dust generatedin HS facilities, the speed with which itis generated, and the inability to breakdown equipment for cleaning make it dif-ficult to effectively sanitize these facilitiesin season, given current conditions andavailable methodology.

Before SanitationD After Sanitation

7

:Presumptive E. co/i

682 FOOD PROTECTION TRENDS I SEPTEMBER 2007

ACKNOWLEDGMENTSThis rcscarch \VaS sUp1)Ortcd b y the

Almond Board of California. We aregrateful for the cooperation of membersof the almond industry. Thanks arcextended to Charm Sciences Inc. forloan of the ATP_biolumifleScCncc system(Charm LUMinator T) for this study.Special thanks are given to Steve Maysfrom Davis, CA, for providing technicalsupport in preparing surface samples forthis study, and to Aaron Uesugi for sup-port during HS facility swabbing. Thanksalso go to Sylvia Yada for her assistance inediting this paper.

REFERENCESI. Almond Hullers and Processors

Association. 2004. Huller/shellergood manufacturing practices andsanitation manual. Available at:http://www.ahpa.net/upl0ad/d0c5/ahpasant_man.pdf. Accessed 17January 2007.

2. Code of Federal Regulations. 2003.21 CFR 110—Current good manu-facturing practice in manufacturing,packing or holding human food.

Section I 10.19 Exclusions.Availableat: http://www.cfsan.fda.gov/lrd/cfr I I 0.html. Accessed 20 January2007.

3. Du, W.-X., and L. J . Harris. 2005.Evaluation of the efficacy of aque-ous and alcohol-based quaternaryammonium sanitizers for reducingSalmonella in dusts generated inalmond hulling and shelling facilities.Poster presented at Institute ofFood Technologists Annual Meet-ing, New Orleans, LA, 16 - 20 July2005.

4. Evancho, G. M., W. H. Sveum, L. J.Moberg, and J . F Frank. 2001. Mi-crobiological monitoring of foodprocessing environments, p.25-35.In F. P. Bownes and K. A. Ito (ed.),Compendium of methods for themicrobiological examination offoods, 4th edition. American PublicHealth Association. Washington,D.C.

5. Isaacs, S., J . Aramini, B. Ceibin,J.A. Farrar, R.Ahmed, D. Middleton,A. U.Chandran, LJ.Harris, M. Howes,E. Chan, A. S. Pichette, K. Campbell,A. Gupta, L. Y. Lior, M. Pearce,C. Clark, F. Rodgers, F Jamieson,

I. Brophy, and A. Ellis. 2005. Aninternational outbreak of sal-monellosis associated with rawalmonds contaminated with a rarephage type of Salmonella Enteritidis.J. Food Prot. 68:191-198.

6. Patterson,J.T. 1971. Microbiologi-cal assessment of surfaces. J . FoodTechnol. 6:63-72.

7. Scheusner, D. L. 1982. Methods toevaluate cleaners and sanitizers.J . Food Prot. 45:1257-1260.

8. Troller, J . A. 1993. Sanitation in foodprocessing, 2nd edition, p. 48-49.Academic Press, San Diego, CA.

9. Umland, G. A., A. J . Johnson, andC. Santucci. 2003. Cereal foodplant sanitation, p. 443-455. In Y. H.Hui, L. B. Bruinsma, J . R. Gorham,W.-K. Nip, P. S. Tong, P. Ventresca(ed.), Food plant sanitation. MarcelDekker. NY.

10. U.S. Environmental ProtectionAgency. 1995. AP 42, Fifth Edition

Compilation of air pollutantemission factors,Volume I: Section9.10.2.1 Almond processing. Avail-able at: http://www.epa.gov/ttn/chief/ap42/609/fmnal/c9s I 0-2a.pdf.Accessed 20 January 2007.

3-A SSI Announces Updated 3-A Sanitary Standards

3-A Sanitary Standards, Inc. (3-A SSl) announces the availability of two updated3-A Sanitary Standards.

The revised 3-A Sanitary Standards include:

Standard 84-02 Personnel Access Ports for Wet Applications —This standard coversthe sanitary aspects of atmospheric or pressure-type personnel access ports andassociated covers for wet applications.Standard 20-25 Multiple-Use Plastic Materials —This sanitary standard covers thematerial requirements of plastics for multiple-use as product contact and/or cleaningsolution contact surfaces in equipment for production, processing, and handling of milkand milk product(s).Test criteria are provided for plastics as a means of determiningtheir acceptance as to their ability to be cleaned and to receive effective bactericidaltreatment and to maintain their essential functional properties and surface finish inaccelerated use-simulating tests.

Copies of the documents are now available for purchase in electronic format or printedversion through the 3-A SSI Web site at www.3-a.org.

SEPTEMBER 20071 FOOD PROTECTION TRENDS 683