Residence Time of Products in Holding Tubes of High-Temperature Short-Time Pasteurizers

Download as pdf or txt
Download as pdf or txt
You are on page 1of 6

Residence Time of Milk Products in Holding Tubes of

High-Temperature Short-Time Pasteurizers


R. W. DICKERSON, JR., A. M. SCALZO, R. B. READ, JR., and R. W. PARKER
National Center for Urban and Industrial Health
U. S. Public Health Service, Cincinnati, Ohio

Abstract water in turbulent flow, Jordan et al. (3) in-


Holding tubes are sized by performing jected tracers o£ salt, dye, and bacteria in hold-
the salt conductivity test on water flow- ing tubes and found no difference in residence
ing at pasteurization temperature. Because time measurements by the three methods. This
dairy products have a higher viscosity than indicates that the saline tracer is an adequate
water, the fastest product particles could method for measuring residence times of bac-
be held for shorter times than predicted by teria in turbulent flow through the holding tubes
the salt test. A study was done to deter- of H T S T pasteurizers. The works cited previ-
mine the error attributable to this practice. ously deal exclusively with water, but no differ-
A high-temperature short-time pasteurizer ence in the residence times of water and milk
unit was operated at 1,130 liters/hour at was expected (4).
pasteurization temperature, and residence Products such as cream, ice cream mix, and
times were determined by measuring the condensed milk are expected to have a residence
time required for an injection of radio- time different from that o£ water because of
active iodine to traverse the holding tube. two different types of flow (6), laminar and
Residence times of milk products were 21.9 turbulent. F o r a given set of conditions, the
and 23.5 sec, respectively, for 16.5 and type of flow can be predicted from the dimen-
10% milk-fat ice cream mix, 22.7 see sionless parameter (5) called Reynolds number
for cream (40% milk f a t ) , and 17.2 sec NR~ = p V..,~ D/t~ Equation 1
for condensed skimmilk (40% solids) flow-
ing in a tube sized for 26 sec by salt injec- W i t h water flowing through a holding tube (in-
tion; and 15.8 and 15.5 see, respec- side diameter ~ 3.56 era) at a rate of 1,200
tively, for milk and chocolate milk flowing liters/hour and a temperature of 80 C, the
in a 16-sec tube. Sizing holding tubes Reynolds number is 26,000, the flow is turbu-
by the salt test is satisfactory for milk and lent, and the shortest residence time is 18% less
chocolate milk; however, to ensure proper than the average residence time (5). Under the
holding times for ice cream mix, cream, and same conditions, with ice cream mix instead of
condensed skimmilk, tubes should be sized water, the Reynolds number is 800, the flow is
by calculating fastest particle velocity of laminar, and the shortest residence time is 50%
product in laminar flow. less than the average residence time (5). There-
fore, when an H T S T pasteurizer is timed for an
When milk flows through the holding tube of ice cream mix standard of 25 sec using
a high-temperature short-time ( H T S T ) pasteur- water, the actual residence time of the fastest
izex, the milk particles move at different veloci- particle of ice cream mix will be less than 25
ties and, therefore, have different holding times. sec, because of the change in flow profile
The particles at the center of the tube flow the from turbulent flow with water during salt tim-
fastest (shortest holding time) and the particles ing to laminar flow with ice cream mix.
near the tube wall move slowly (longest holding The objective of this work was to compare
time). Jordan and March (4) demonstrated the residence times of milk products in the hold-
that the difference between fastest and slowest ing tube of a I t T S T pasteurizer with that of
particles can be as much as 10 sec in a 15- water as determined by the salt injection tech-
sec holding tube. Because of this, the length nique. Because of the high electrical conduc-
of the holding tube is determined by operating tivity of milk products, attempts to use the salt
the pasteurizer with water, injecting a saturated injection technique with product were not suc-
salt solution in the holding tube, and measuring cessful and, therefore, radioactive tracers were
the time required for the injection to traverse used to measure residence time of the milk prod-
the tube (7). The prime concern, however, is ucts. Measurements were made on cream, con-
not the residence time of a salt solution, but densed skimmilk, chocolate milk, raw milk, and
rather the residence time of pathogenic bacteria two ice cream mixes o£ different milk f a t con-
which may be in the raw product. Working with tent.
1731
1732 DICKERSON, SCALZO, READ, A N D PARKER

Materials and Methods salt injection technique when the pasteurizer


Products. All products were obtained from was operated with water at a temperature of
the Cincinnati market. The ice cream mix for- 80 C and nominal capacity. F o r tests on choco-
mulations were supplied by A. B. Jones of late milk and raw milk, holding tube length was
French-Bauer, Inc., 1 Cincinnati. The products changed to obtain a minimum residence time of
were made from raw ingredients, except for the 16 sec. Sizing of the holding tube by the
heat processes used to manufacture the cream salt injection technique with water was per-
and condensed skimmilk used in the ice cream formed at a temperature of 75 C for chocolate
mixes. The product formulations were as fol- milk and 73 C for raw milk. F o r ice cream mix,
lows : cream, and condensed milk, the holding tube
Ice cream mix: 16% milk fat containing was 9.2 m long and for chocolate milk and raw
24% liquid sugar (corn and cane), 23% con- milk, 6.1 m. One test was performed on ice
densed skimmilk (30 % total solids), 40 % cream cream mix in a 4.76-cm-inside-diameter tube
(40% milk f a t ) , 12.7% raw milk (3.7% milk that was 5.7 m long to evaluate the effect of
f a t ) , and 0.25% gum stabilizer. different tube diameters on residence times.
Ice cream mix: 10% milk fat, containing Preparation of tracers. Three radioactive
22% liquid sugar (corn and cane), 34% con- tracers were used for measuring residence time
densed skimmilk (30% total solids), 24% cream of products: 1~I in solution, silver iodide crys-
(40% milk f a t ) , 19.7% raw milk (3.7% f a t ) , tals, and bacteria stained with iodine. The for-
and 0.3% gum stabilizer. mer was made by adding 100 t,c of 1~I in I ml of
Cream: 40% milk fat. distilled water to 50 ml of product. Radioactive
Condensed skimmilk : 40% total solids. silver iodide crystals were made by adding 100
Chocolate milk: 3.5% milk fat, containing /~e of laI to K I in solution and mixing with
1% cocoa, 9% liquid cane sugar, and 90% raw AgNO3 to form the radioactive precipitate. The
milk (3.7% f a t ) . precipitate was suspended in 50 ml of product
Raw milk: 4.0% fat. and used as the tracer material. To prepare
Pasteurizer. The H T S T system (Fig. 1) had bacteria labeled with z~I, Staphylococcus aureus
a nominal capacity of 1,130 liters/hour. The in- was grown overnight at 35 C in brain heart in-
side diameter of the holding tube was 3.56 cm fusion broth agitated on a rotary shaker set for
and the length was chosen to obtain a minimum 150 times per minute. The bacteria were har-
residence time of 26 sec, as measured by the vested by centrifugation at a relative centrif-
ugal force of 6,000 for 20 rain, washed in
Mention of commercial products does not imply distilled water, and stained by adding an equal
endorsement by the U.8. Public Health Service. volume of crystal violet to the packed cells in

FLOW-DIVERSIVALVE
ON - - GAMMASCINTILLATIDETECTOR
ON
l ~ RECORDER- CONTROLLER
TEMPERATURE
SENSOR

FORWARD
FLOW
UNE~~ E I I v l I y I I I I I D D E E

, ~LSAMPLING
GAMMA
SCINTILLATIONcooLER
DETECTOR-R~EGENER HEATER HOLDINTUBE
G )POR1
SECTION SECTION l I SECTION
COLW
DATEIN--I
R 7PI°tes~ ~ [I ~°te ~ -HOW
TATEIN
R
PRODUCT
TOSTORAGE |/~// In--'~V~'~L~ V I'

RAWPRODUCT
INLET
~ . E ~( ,TIi UGALBOOSTER
PUMP' ~METER,L
PUMP
RAWPRODUCT
SUPPLY
TANK
:Fze 1. High-temperature short-time pasteurizer showing holding tube and gamma scintilla-
tion detectors.
J. DAIRY SCIENC~ VOL. 51, NO. 11
PRODUCT RESIDENCE TIMES 1733

each centrifuge bottle. After 1 min of stain- tainer was mounted in a water bath, and tem-
ing, the packed cells were diluted to 100 ml perature was controlled at the pasteurization
with distilled water and resuspended. After temperature. Measurements of viscosity and
centrifugation and removal of the supernatant density were completed within 3 min after
fluid, a volume of grams iodide equal to the drawing the sample and temperature of the sam-
volume of packed cells and supplemented with ple was within 0.5 C of pasteurization tempera-
100 ~c of 1311 was added and allowed to react tare. Viscosity was determined by means of a
with cells for 1 rain. This was followed by Brookfield LVT viscometer without the spindle
seven washing steps sin~ilar to those just de- guard and with the container filled to within
scribed, to wash out free radioactive iodine. 2.5 cm of the top. A measurement of the vis-
Product was added to the stained bacteria to cosity of a calibrated oil indicated an accuracy
yield an injection volume of 50 ml and the in- of better than 5%. Density was measured with
jection contained approximately 10TM bacteria. a Mettler balance (modified for under-the-pan
All tracer-product mixtures were heated so that weighing) to determine the weight of a plum-
temperature at injection averaged 80 C. A sat- met (from a Westphal balance) submerged in
urated salt solution was used to measure resi- the product.
dence time with water. Nuclear instrumentation. Two gamma-sensitive
Test procedure. About 1,500 liters of raw scintillation detectors (Model no. 821330) were
product were used for each test, and this yielded coupled with a dual-channel pulse-height ana-
80 min of running time. No attempts were lyzer (Model no. 8742) and a dual-channel rate-
made to recirculate pasteurized product. The meter (Model no. 8732), all manufactured by
product was normally used the same day deliv- Nuclear-Chicago Corp. The two signals from
ered, and always within two days. the ratemeter were recorded with a multi-
The radioactive tracers were placed in 50-ml channel light beam oscillograph (Honeywell,
glass bottles (without product) and each bottle Model no. 1508). The scintillation detectors
placed in the holding tube under the scintilla- were mounted outside the holding tube, and the
tion detector and the activity measured through radioactive tracer was detected through the
the wall (0.125-cm thick) of the stainless steel stainless steel tube wall. The anticipatory up-
holding tube to determine that it was 45,000 stream detection of the tracer was prevented by
counts/minute or greater. Product was then lead-shielded collimating nosepieces mounted on
added to the tracers and the glass bottles placed each detector; the field of view was less than
in the water bath and heated to injection tem- 8 cm of holding tube length.
perature. The pulse height analyzer was adjusted to
The pasteurizer was operated with water and c o u n t d i s i n t e g r a t i o n s h a v i n g e n e r g y levels
adjusted to the pasteurization temperature. greater than 0.344 Mev and less than 0.384 Mev
Pumping rate was adjusted to the rated capacity
of 1,130 liters/hour and residence time in the TABLE 1. Comparison of the residence time of a
holding tube measured by the salt injection tech- radioactive solution, radioactive-labeled bac-
nique (7). While the pasteurizer was operating teria, and radioactive silver iodide in
with water, flow was switched to product, and ice cream mix (16% milk fat). ~
pumping rate was readjusted to compensate for Injec- Rela-
the differences between water and product ac- tion tire Res-
cording to the requirements of the 1965 Pas- tempera- activ- idence
teurized Milk Ordinance (7). Tracer ture ityb time
The radioactive tracer was transferred from (counts/
the water bath to a 50-ml syringe, and tempera- (C) rain) (see)
t-are of the tracer was measured prior to injec- 1~I solution 80 70,000 21.9
tion. The tracer was injected at the beginning Bacteria labeled
of the holding tube through the same probe used with 181I 85 48,000 22.4
Radioactive silver
for salt timing. All measurements of residence iodide 80 58,000 22.4
time of products were performed in duplicate, 1~I solution 25 66,000 22.1
and results differed by less than 0.5 sec. l~low rate was 1,130 liters/hour, and the hold-
Viscosity and density measurements. A sam- ing tube sized for 26 see as measured by the salt
ple of product was taken from the holding tube test using water. The pasteurizer was operated
during operation for measurement of viscosity at 80 C. The inside diameter of the holding tube
and density measured at pasteurization temper- was 3.56 cm and its length 9.2 m.
b The tracer was placed in the holding tube,
ature. The sample was placed in a container under the detector, and activity recorded before
4.76 cm in diameter and 22.9 cm long. The con- the tracer was mixed with 50 ml of product.
J. DAIRY SCIENCE, VOL. 51, NO. 11
1734 D I C K E R S O N , SCALZO, READ, AND P A R K E R

(the energy peak for 1~I). This yielded a back- 1~I solution were heated to pasteurization tem-
ground count of less than 12 counts~minute and perature before injection into the holding tube
facilitated interpretation of the oscillograms. and the temperature of a second tracer of ~ I
Response time of the ratemeter was adjusted solution was 25 C prior to injection. All four
to 0.5 sec and the frequency response of the tracers had residence times between 21.9 and
oscillograph galvanometers was 120 cycles/sec- 22.4 sec and, since the difference between
ond. Because of the fast response time (0.5 them is negligible, any of the tracers could have
see), background counts were recorded as been used. All subsequent tests were performed
individual pulses spaced about 5 sec apart. with the hot (80 C) solution of ~ I in product.
The presence of the tracer under the detector The length of the holding tube was determined
was manifested by an abrupt increase in fre- by operating the pasteurizer with water and in-
quency of counts on the oscillogram. The oscil- jecting a saline tracer into the holding tube (7).
lograph paper speed was 1 cm/second, and ac- W i t h a flow rate of 1,130 liters/hour of water,
curacy of the time measurements taken from the average of six measurements was 26.3 see
the oscillogram was within 1%. and the individual measurements were
Detection of the fastest particle. The instru- within 0.5 see of each other. Using this
mentation used in both the salt and radioactive same holding tube at the same flow rate, the
tracer techniques is limited by a threshold sensi- residence times of other milk products were
tivity. Because of this, we did not detect the measured (Table 2). These products traversed
fastest particle. Actually, it is not necessary to the holding tube in times as brief as 17.2 sec.
detect the fastest particle only the fastest mi- Condensed skimmilk exhibited the short-
croorganism. A radioactive dose of about 20 mc est residence time and differed from fully de-
per tracer would be required, however, to detect veloped laminar flow by less than 8%.
the fastest radioactively labeled bacterium, and A shorter holding tube was used for raw milk
we believed the data would not be worth the and chocolate milk. W i t h a flow rate of 1,130
extensive decontamination controls and proce- liters/hour of water, the average of six resi-
dures required for the large doses, especially dence time measurements by the salt test was
since the residence times of the first tracer par- 16.6 sec. Using the ~1I tracer, residence
ticle and the first 1% of the tracer particles time was 15.8 and 15.5 sec, respectively, for
differ by only 1% in fully developed laminar raw milk and chocolate milk. The difference be-
flow. This can be demonstrated by the equation tween the residence time of salt in water and
for the velocity profile of laminar flow (5) the residence time of milk was 0.9 sec.
V -~ 2 V~,~ [1 - (r/R) ~] Equation 2
Discussion
The first 1% of the tracer occupies a volume During these tests, ice cream mix was held at
parabolic in shape, with its center located at pasteurization temperature for only 22 see
the tube centerline and its edge located 0.1 R instead of the 26 see indicated by the salt
away from the tube centerline. Equation 2 in- test. The error was 17%. F o r condensed skim-
dicates the velocities and, therefore, residence milk, the error in residence time predicted by
times of the particle at the tube centerline and the salt test was 33%. With the worst possible
particle at a radial distance 0.1 /~ away from condition (large diameter tube with no bends
the tube centerline differ by 1%. and a highly viscous product in laminar flow)
the error in residence time, predicted by the
Results salt test, would be 45%. Because of these er-
Aiba and Sonoyama (1) have shown that, in rors, a new technique is needed for sizing 25-
laminar flow, the residence time of a tracer in sec holding tubes.
solution is different from that of a tracer in No method is currently available for measur-
suspension. Using radioactively labeled bac- ing the residence time of milk products in the
teria and ~P in solution, they demonstrated field. The radioactive tracer method is not rec-
that the residence time of bacteria was the same ommended, because of the high cost of instru-
as that predicted in fully developed laminar mentation and the loss of contaminated product.
flow ; whereas, the residence time of the ~P solu- Thermal methods of measuring holding time are
tion was less than that of the bacteria. Because of interest, but these require correction factors
of this, we measured the residence time of bac- dependent on type of equipment, volume of in-
teria labeled with ~31I, radioactive silver iodide, jection, and method of obtaining the thermal
and solutions of mI in ice cream mix to detect change in the holding tube (2).
any significant differences between tracers. The data shown here cannot be used in a gen-
The bacteria, silver iodide, and one tracer of eral application because the residence time of
J. /).~]~R~ SCIenCE ~/oL. 51, NO. 11
PRODUCT RESIDENCE TIMES 1735

TABLE 2. Comparison of holding time predicted by the salt test with measured residence time of radio-
active tracer in product.
Holding Error in
time Measured predicting
predicted residence product
by salt time of residence
Product Product test on ~ I in time by the
Product viscosity ~ density~ water productb salt test
(CP) (g/ee) --(see) (%)
Condensed skimmilk (40% solids) ~ 17~ 1.15 25.6 17.2 33
Ice cream mix (16% milk fat) ~ 15~ 1.06 26.3 21.9 17
Ice cream mix (10% milk fat) ~ 11~ 1.08 26.3 23.5 11
Cream (40% milk fat) ¢ 4~ 0.96 26.3 22.7 14
Chocolate milkd 1.2h 1.03 16.6 15.5 7
Raw milk~ 1.0h 1.01 16.6 15.8 5
Measured at pasteurization temperature.
b Tracer heated to 88 C before injection.
° Pasteurizer operated at 80 C, with holding tube 9.2 m long.
d Pasteurizer operated at 75 C, with holding tube 6.1 m long.
e Pasteurizer operated at 73 C, with holding tube 6.1 m long.
f Brooktield viscometer, no. 1 spindle, 60 rpm.
Brookfield viscometer, UL adapter, 30 rpm.
h Brookfield viseometer, UL adapter, 60 rpm.

milk products in holding tubes is dependent on lengthened to meet the 25-see requirement.
tube diameter, tube length, number of bends in Rather than increase the length of every hold-
the holding tube, and viscosity of the product, ing tube, it appears more reasonable to change
We performed a test on a holding tube with an the pasteurization standard to one that will con-
inside diameter of 4.76 cm and the residence form to the residence time now available with
time of ice cream mix (16% milk fat) was 20.6 existing holding tubes. A 15-see residence
sec. Increasing the size of the holding tube time, computed with Equation 6, is available in
from 3.56 to 4.76 em decreased holding time presently installed holding tubes. The required
from 21.9 to 20.6 see. temperature may be determined by calculating
By assuming fully developed laminar flow the z-value (the temperature increase that per-
(the worst condition), the length of the holding mits a tenfold reduction in holding time while
tube required for pasteurization can be accu- maintaining the same lethality of the process)
rarely determined by calculation. This is accom- between the two equivalent processes, 68.3 C
plished by measuring the maximum pumping for 30 rain and 79.4 C for 25 see, and
rate of the pasteurizer and calculating the hold- using this value (6.0 C) to calculate the equiv-
ing tube length using the equation of continuity aleut temperature of a 15-see holding time.
V,~ = Q/A Equation 3 The new standard is 81.1 C for 15 see.
To implement the suggested change, it is nec-
Furthermore, from Equation 2 essary only to raise the diversion setting of the
Vm,x = 2 V,~ Equation 4 recorder-controller from 79.4 to 81.1 C. The
and finally, holding tube length is, calculation of holding tube length (Equation 6)
L = Vm,x t Equation 5 is a simple one and the measurement of the
Substituting Equations 3 and 4 into Equation 5 residence time of a saline tracer in water is
yields not needed.
L = 8 Qt/~D 2 Equation 6 The improvement is necessary only for 25-
sec holding tubes. The residence time pre-
Equation 6 may be used to calculate the re- dieted by the salt test in a 15-see holding
quired length of holding tubes, or if the length tube is only 1 sec longer than the actual
of the holding tube is fixed, the equation may residence time of product and, since differences
be used to calculate maximum pumping rate, Q. of 0.5 see are permitted between individual
The precision of the method is the same as that tests using the salt tracer, a similar Change for
of the measurement of pumping rate--usually 15-see holding tubes is not justified.
within 2%.
I f the above technique is employed with a References
pasteurization standard of 79.4 C for 25 see, (1) Aiba, S., and T. Sonoyama. 1965. l~esidenee
most holding tubes would have to be time distribution of a microbial suspension
J. DAI~Y SCIeNCe., VOL. 51, NO. I 1
1736 DICKERSON, SCALZO, READ, AND PARKER

in a straight pipe. J. Fermentation Technol., Nomenclature


43 : 534. .A, cross-sectional area of the holding tube
(2) Jordan, W. K., and R. F. Holland. 1953. D, inside diameter of the holding tube
Studies on thermal methods of measuring
the holding time in high-temperature short-
L, length of the holding tube
time pasteurizers. J. Milk and Food Teeh- Mev, million electron-volts
nol., 16: 15. NR~, Reynolds number
(3) Jordan, W. K., R. F. Holland, and J. C. Q, volumetric flow through holding tube
White. 1949. The determination of the r, radial distance from centerline of holding tube
holding time in hlgh-temperature, short-
time pasteurizing units. J. Milk and Food R, inside radius of holding tube
Teehnol., 12 : 87. t, holding time required for pasteurization
(4) Jordan, W. K., and R. P. March. 1953. Studies V, velocity of particle in holding tube
on overholding in high-temperature short- Va,a, average velocity of all particles in holding
time pasteurizers operated on water. J. tube
Dairy Sei., 36 : 614.
V~,,,,, velocity of fastest particle in a holding tube
(5) Knudsen, J. G., and D. L. Ka.tz. 1958. Fluid
during laminar flow
Dynamics and Heat Transfer. McGraw-
t*, absolute viscosity
Hill, New York. p. 149.
p, density
(6) Reynolds, O. 1883. A a experimental investi-
gation of the circumstances which determine
whether the motion of water shall be direct
Acknowledgments
or sinuous, and of the law of resistance of The authors acknowledge the helpful suggestions
parallel channels. Roy. Soc. London, Phil. of J. E. Campbell, who recommended radioactive
Trans., 174: 935. silver iodide as a tracer, and the assistance of
(7) U.S. Public Health Service. 1965. Grade J. Bradshaw in the bacterial preparations. We
"A" Pasteurized Milk Ordinance--1965. are particularly grateful for the assistance of
Recommendations of the U.S. Public Health A. B. Jones of French-Bauer, Cincinnati, Ohio,
Service. U.S. Government Printing Office, who formulated and prepared the milk products
Washington, D . C . p . 163. used.

J. DAI~Y ScIBI~e~ Voi~. 51, NO. ii

You might also like