Case Study: Fine Mist Disinfection on a Beverage Bottle Filler

A beverage filler disinfects three surfaces before the product arrives: the bottle, the closure, and the filler valve. Each runs on a different clock. The bottle crosses the station in under a second, the cap waits in a chute for minutes, and the valve runs all shift.

This case describes a European beverage bottling plant that replaced flood rinsing and soak baths with fixed point fine mist. Every figure below is a working value for a line of this class.

Case Snapshot

The reference configuration is the generic shape of the installation.

Item	Reference configuration
Industry	Beverage bottling, still and carbonated soft drinks in PET
Duty	Bottle disinfection, closure disinfection and filler valve disinfection ahead of filling
Equipment	Rotary filler at 24,000 bottles per hour, 500 mL PET bottle, 65 mm body, 80 mm pitch, 32 m/min conveyor
Original setup	Inverted bottle flood at 6.0 mL per bottle, a 20 mL potable rinse, a closure soak bath, and manual valve foaming each shift
Nozzle chosen	External mix air atomising misting nozzles at fixed points, gated on bottle position, 30 degree fans
Operating point	2.4 L/h per nozzle at 1.5 bar liquid and 2 bar atomising air, 0.8 s spray window, 1.2 mL per bottle
Material requirement	316L body, PTFE seals, PVDF wetted parts, FDA 21 CFR 177 materials, no copper or zinc
Compliance	EHEDG cleanable design, CIP compatible to 80 °C, 0.5 ppm reference rinse residual, documented spray settings

The output class matters more than the drink. A rotary filler between 18,000 and 30,000 bottles an hour is the common shape. The sizing transfers to a linear filler or a can line.

The Challenge

The filler ran 24,000 bottles an hour, which is 400 bottles a minute. Bottles ran at an 80 mm pitch, so the conveyor moved at 32 m/min, or 0.53 m/s. At 2,500 running hours a year the plant filled 60 million bottles.

Bottle disinfection was a flood. Bottles were inverted over a manifold of 24 spray nozzles that washed the inside with 6.0 mL of peracetic acid solution at 2,000 ppm. A potable bottle rinsing stage followed at 20 mL per bottle. Closures soaked in a shared bath, and the filler valves were foamed by hand each shift.

Volume was the first problem. At 400 bottles a minute the flood manifold passed 144 L/h, which is 360,000 L of diluted peracetic acid a year. The rinse added 1,200 m³ of water.

Contact was the second. A flooded bottle drains in about 1.5 s, so the flood gave a Ct of roughly 3,000 ppm seconds. The film ran to the base while the neck ring and the thread stayed wet but thin.

The soak bath was the third. Closures sat in a tank dosed by hand at 1,500 to 2,000 ppm, and the strength drifted about plus or minus 30 percent a shift.

The valves were the fourth. Manual foaming took two operators 35 minutes per shift, so 250 shifts a year cost the line 146 hours.

The last problem was coverage. A flood wets the path it hits, so the shadow behind the bottle profile and the gap between the valve bodies stay dry.

The Solution

The plant stopped treating disinfection as a wash and started treating it as a metered mist at fixed points. Four numbers set the design: 1.2 mL of solution per bottle, 1,500 ppm, a 0.8 s window, and a 34 s dwell.

The dose follows from the surface. A 500 mL PET bottle carries about 0.043 m² of inner wall. A 1.2 mL dose there is a film 28 µm deep, which holds wet contact across the wall and still dries before the filler.

At 400 bottles a minute, 1.2 mL per bottle is 480 mL/min, or 28.8 L/h. Twelve misting nozzles in two manifold rows share that load at 2.4 L/h each. Flow is set per metre of conveyor, so the dose holds when the line slows.

Drop size was the second decision. Air atomising nozzles at 2 bar air and 1.5 bar liquid make a Sauter mean diameter of 50 to 100 µm at this flow. A hydraulic tip at the same 2.4 L/h makes 150 to 250 µm drops and throws them past the bottle.

Layout was the third. Four nozzles aim down into the mouth and eight aim at the shoulder and body, four per side. Each is a 30 degree fan at 170 mm standoff, which covers 91 mm against a 65 mm bottle.

Timing was the fourth. A photoeye gates the valve bank, and a drip free tip with a spring loaded needle closes in 60 ms. The window is 0.8 s, the time one bottle spends in the 0.45 m zone at full speed.

The plant also dropped the closure bath and the manual valve foam. This filler disinfection layout mists 0.4 mL per closure on its way to the capper and keeps the valve bodies wet at 6 L/h.

Control of the peracetic acid loop was the last change. A dosing pump holds the manifold at 1,500 ppm plus or minus 50 ppm. A 3.0 mL mist rinse per bottle brings the wall residual below the 0.5 ppm reference limit.

The Engineering Detail

The Line Runs 24,000 Bottles an Hour, So the Window Is 0.8 s

Line speed is an output divided by a pitch. At 24,000 bottles an hour the filler takes 400 bottles a minute. A 65 mm bottle on an 80 mm pitch puts the conveyor at 400 x 0.080 = 32 m/min, or 0.53 m/s.

The spray station is 0.45 m long, which is what twelve nozzles in two rows occupy. A bottle crossing it at 0.53 m/s is inside for 0.85 s, so the window is set at 0.8 s.

Residence is not contact time. The wet film stays on the wall from the station to the rinse, a run of 18 m or 34 s.

The dose follows the geometry. A bottle needs 1.2 mL and bottles arrive every 0.15 s, so the bank delivers 8 mL/s. Across twelve nozzles that is 0.67 mL/s each, or 2.4 L/h.

The control law is written per metre of conveyor. At 0.53 m/s and 8 mL/s the bank lays down 15 mL/m, which is 1.2 mL per bottle at an 80 mm pitch. Below full speed the bank pulses at 5 Hz to hold the same 15 mL/m.

Step	Value	Arithmetic
Bottles per hour	24,000	given output
Bottles per minute	400	24,000 / 60
Bottle pitch	80 mm	65 mm body plus a 15 mm gap
Line speed	32 m/min	400 x 0.080 m
Line speed	0.53 m/s	32 / 60
Zone length	0.45 m	twelve nozzles in two rows
Residence in the zone	0.85 s	0.45 / 0.53
Spray window	0.8 s	set by the photoeye gate
Solution per bottle	1.2 mL	dose target
Manifold flow at full speed	28.8 L/h	24,000 x 1.2 mL
Flow per nozzle, twelve nozzles	2.4 L/h	28.8 / 12
Flow per metre of conveyor	15 mL/m	8 mL/s / 0.53 m/s
Dwell to the rinse	34 s	18 m / 0.53 m/s
Rinse per bottle	3.0 mL	residual target

Drop Size at 50 to 100 µm and the Shadow the Bottle Casts

A mist only disinfects what it touches, so the drop has to reach the neck ring. The duty sits at a Sauter mean diameter of 50 to 100 µm, a working band for air atomising misting nozzles at 2.4 L/h.

Count the drops. At a 75 µm SMD one drop is 2.2 x 10⁻⁷ mL, so 1.2 mL is about 5.4 million drops. Over 0.043 m² that is 126 drops per square millimetre, the difference between a wet wall and a running film.

The air stream leaves the cap at 20 to 40 m/s and slows to 3 to 8 m/s at 170 mm. A 75 µm drop carries a Stokes number near 1.3, so it crosses to the wall.

The shadow is the design problem. A single aim leaves a dry band behind the profile of about 12 percent of the surface. A second aim from the opposite side and a nozzle pointed into the mouth close it.

Beverage disinfection is not classified by the ASABE S572.1 or ISO 10625 colour charts, which cover agricultural fans. The same laser diffraction method still applies.

A hydraulic tip at the same 2.4 L/h needs a 0.15 mm orifice and about 3 bar, and it makes 150 to 250 µm drops. That orifice will not pass the coarse fraction a recirculated peracetic acid line carries.

The K-Factor at 1.5 bar and Why Q Follows the Square Root of Pressure

A nozzle does not have one flow rate, it has a K-factor. The relation is Q = K x sqrt(P), with Q in L/h and P in bar. The chosen tip has K = 1.96, so at 1.5 bar it passes 1.96 x 1.22 = 2.4 L/h.

Pressure is a weak lever on volume. Doubling it from 1.0 to 2.0 bar raises flow by 41 percent, not 100 percent, and the drops fall 15 to 20 percent finer.

Pressure	Flow per nozzle	Arithmetic
1.0 bar	1.96 L/h	1.96 x sqrt(1.0)
1.5 bar	2.40 L/h	1.96 x sqrt(1.5), chosen
2.0 bar	2.77 L/h	1.96 x sqrt(2.0)
2.5 bar	3.10 L/h	1.96 x sqrt(2.5)

Hold the pressure and gate the time instead. The plant fixes the liquid at 1.5 bar and modulates duty on a conveyor encoder. Drop size then holds from 6,000 to 24,000 bottles an hour.

Size the header on the flow while the bank is open. At 6,000 bottles an hour the bank runs a 25 percent duty, and a 6 mm bore line carries 28.8 L/h at 0.28 m/s.

Coverage Arithmetic: Flow x 60 / Area

Coverage reads coverage = flow x 60 / area, with flow in L/min and area in m² to give litres per hour per square metre. The manifold passes 0.48 L/min over 0.135 m² of tunnel cross section, so coverage is 0.48 x 60 / 0.135 = 213 L/h per m².

Per bottle the number is simpler. The 0.043 m² of inner wall takes 1.2 mL, which is 28 mL/m² and a film 28 µm deep. The old flood put 6.0 mL on the same area, a film 140 µm deep that runs to the base.

Width follows the standoff. For a flat fan, width = 2 x standoff x tan(angle / 2). At 30 degrees and 170 mm that is 2 x 170 x tan 15 = 91 mm.

Standoff	Covered width	Relative deposit	Note
120 mm	64 mm	1.42	fan under-covers the 65 mm body, bare neck ring at both edges
170 mm	91 mm	1.00	chosen: 13 mm of margin per side
200 mm	107 mm	0.85	8 mm of pattern lands past the bottle on each side
300 mm	161 mm	0.57	more than half the pattern is overspray

A 60 degree fan needs only 79 mm of standoff to cover 91 mm, which puts the tip inside the splash zone. A 15 degree fan needs 346 mm, which suits a tunnel but makes alignment critical.

Chemistry: 1,500 ppm Peracetic Acid and the 0.5 ppm Rinse Target

Peracetic acid suits beverage bottle disinfection because it works at low strength and breaks down into acetic acid and oxygen. Typical strength is 0.1 to 0.2 percent, or 1,000 to 2,000 ppm. This plant runs 1,500 ppm.

Kill follows concentration times contact time. That product, Ct, is the figure that carries into a validation. The old flood gave about 1.5 s of contact at 2,000 ppm, a Ct of roughly 3,000 ppm seconds.

The mist gives 0.8 s of application plus a 34 s dwell at 1,500 ppm. Ct is 1,500 x 34.8 = 52,200 ppm seconds, about 17 times the old figure on 80 percent less solution.

That gap is the argument for fine mist. The flood ran off the wall into the base, so much of the solution never touched the surfaces that matter. A 28 µm film stays where it lands.

The residual limit sets the rinse. At a 0.5 ppm reference limit on the wall after rinsing, the rinse has to clear the film left in the neck. The plant mists 3.0 mL of potable water per bottle and titrates the finished residual every two hours.

Metal ions break peracetic acid down, which is why the wetted path carries no copper or zinc.

Concentration drift is the failure that hides. The old bath drifted about plus or minus 30 percent. An inline dosing pump now holds the manifold at 1,500 ppm plus or minus 50 ppm.

Air Atomising Against Hydraulic at 2.4 L/h, and the Drip Free Tip

Air atomising was chosen for a low flow with a fine drop.

Option	Flow per nozzle	Drop size, typical	Note
Hydraulic at 3 bar	2.4 L/h	150 to 250 µm	needs a 0.15 mm orifice, throws past the bottle
Hydraulic at 10 bar	2.4 L/h	90 to 140 µm	0.15 mm orifice, 50 µm filtration, unstable fan edges
Air atomising, internal mix	2.4 L/h	30 to 60 µm	finest drops, chamber holds fluid between pulses
Air atomising, external mix	2.4 L/h	50 to 100 µm	chosen: no chamber, short passage, easy to strip

Internal mix makes the finer drop, and the chamber is the objection. A chamber that holds 5 mL between pulses lets a reactive fluid sit. External mix passes the liquid through a short passage, so very little waits between pulses.

Atomising air is the running cost. Each cap draws about 2 Nm³/h at 2 bar, so twelve draw 24 Nm³/h while the bank is open, or about 60,000 Nm³ a year.

Drip is what the tip has to prevent. A tip without a positive shutoff leaves a hanging drop of about 0.02 mL on the cap face. At 0.53 m/s that drop lands on the next bottle neck. A spring loaded needle shuts in 60 ms and leaves the cap face dry.

316L, EHEDG Cleanability, and Three Failure Modes

The wetted path is 316L stainless for the body and liquid tip, with PTFE seals and PVDF for the air cap and manifold blocks. No copper or zinc bearing alloy sits anywhere in the path.

Elastomer choice is not a detail. Peracetic acid attacks nitrile and degrades FKM, so both are out. PTFE and PVDF are the safe wetted plastics, and EPDM covers the housing seals.

Contact materials meet FDA 21 CFR 177 for repeat use articles, including PTFE at 177.1550 and PVDF at 177.2510. The nozzle follows EHEDG hygienic design: no crevices, drainable surfaces, and a 0.8 µm Ra finish. Wetted parts take CIP at 80 °C, and the unit strips down without tools.

Three failure modes matter on this duty, and each has a tripwire.

A drop that is too large is the first. Above about 200 µm SMD the drop falls rather than turns and pools in the bottle mouth. A bead of 0.1 to 0.3 mL standing in the neck is the symptom, and a check weight on 30 bottles catches it.

A blocked nozzle is the second. A 0.25 mm orifice is 40 percent of the way to blocked by a 100 µm particle. A half blocked tip loses its fan centre and leaves a 60 to 80 mm dry strip in the zone. Filter to 50 µm absolute with a 100 µm guard, and watch header pressure.

A drifting concentration is the third. Hand dosing drifted about plus or minus 30 percent. A fall from 1,500 to 800 ppm cuts the kill by more than 1 log. The loop needs a dosing pump and a two hour titration.

The Results

The table below compares the twelve months after the change with the twelve months before, at the same 60 million bottles.

Metric	Before	After
Peracetic acid solution per bottle	6.0 mL	1.2 mL
Active peracetic acid per bottle	12 mg	1.8 mg
Solution strength	2,000 ppm	1,500 ppm
Solution per year	360,000 L	72,000 L
Active peracetic acid per year	792 kg	119 kg
Rinse water per bottle	20 mL	3.0 mL
Rinse water per year	1,200 m³	180 m³
Effluent COD load per year	984 kg	74 kg
Log reduction on the bottle wall	2.1 log	4.3 log
Residual peracetic acid after the rinse	2.5 ppm	0.4 ppm
Filler stop for manual valve foam	146 h per year	25 h per year
Nozzle strip and clean interval	4 weeks	16 weeks
Compressed air for atomising	none	60,000 Nm³ per year

Peracetic acid solution fell from 6.0 to 1.2 mL per bottle, a cut of 80 percent. Active mass fell further, from 12 to 1.8 mg per bottle, because the mist runs at 1,500 ppm against the old 2,000 ppm.

Rinse water fell from 20 to 3.0 mL per bottle, an 85 percent cut. The stream dropped from 1,200 m³ a year to 180 m³.

COD fell harder than the volume. At 820 mg/L the old stream carried about 984 kg of COD a year. At 410 mg/L the new stream carries about 74 kg, a cut of 92 percent.

The microbiology improved while the chemistry fell. Surface count reduction on the bottle wall moved from 2.1 log to 4.3 log. A 28 µm film that stays on the wall beats a flood that drains to the base.

Manual work fell as far as the chemistry. Valve foaming took two operators 35 minutes a shift and stopped the filler. The nozzle ring keeps the valve bodies wet, so the task is a 6 minute check, 25 hours a year against 146.

Cleaning moved the other way in the good sense. The flood tips were stripped every 4 weeks because scale built on the fan faces. The air atomising tips now run 16 weeks between strips.

One number rose. Atomising air went from nothing to about 60,000 Nm³ a year against a cut of about 670 kg of active peracetic acid.

The compliance audit closed with no major findings. The dose, the window, the concentration and the strip down interval sat on the setting sheet.

Why This Matters

Three moves made this change work, and none of them was the nozzle alone.

The first is to size from the surface, not from the tank. Work out the film depth the wall needs, multiply by the contact area, and get millilitres per unit. Divide by the units a minute, then by the nozzle count.

The second is to hold drop size and gate the volume. Pressure moves flow with a square root and moves drop size with it, so pressure is not a volume control. Fix the pressure and let duty follow the line speed.

The third is to treat contact time as the asset. The mist is weaker than the flood it replaced. It wins because the film stays on the wall for 34 s instead of draining away in 1.5 s.

Three thresholds carry into any beverage line. Keep the Sauter mean diameter between 50 and 100 µm, or the neck ring stays dry. Keep the pattern inside the bottle envelope with 10 to 20 mm of margin. Hold the peracetic acid loop within 50 ppm, or the next validation fails.

Browse the misting nozzle range used here, and read the bottling line spray systems guide for the manifold layout. If your line floods to disinfect, send the bottle pitch, the units an hour and the volume per unit to the BoreJet application team. The bottle disinfection application page shows the same three stations on one schematic.