Instrumentation

656

Laboratory
deodorization:
An overview
of past and
present
equipment
and practical
uses

Process changes in any

stage of the refining of edible oil can have dramatic
effects on finished oil quality. Shifts in consumer preferences for attributes of edible oils are forcing refineries
to re-evaluate processing
protocols and equipment as
well as to use more active
bleaching earths. Many
research and development
edible oil chemists are evaluating the effects these
changes may have on finished oil quality.
Although deodorization is critical to the
evaluation of finished oil quality, many projects stop short of the deodorization step
because equipment, materials, and time are
unavailable or costly.
This paper presents laboratory deodorization options, defines laboratory-scale
deodorization, describes its practical uses,
and provides an overview of past and present deodorization techniques and equipment. In addition, data are presented from
a newly designed six-port microdeodorizer.

This article is by David D. Brooks of

Oil Dri Corp. and Gary R. List of the
USDA National Center for
Agricultural Utilization Research.

Diagram 1.
Microdeodorizer schematic
C
E

Figure 2.

Diagram 2.
Reaction flask setup

J
SPARGER
TUBE

R
N
O

DISTILLATION
MANIFOLD

REACTION
FLASK

THERMOMETER

Volume 13 August 2002 inform

A.Vacuum Pump
B.T-Valve
C. Hg Vacuum Gauge
D. Bleed Valve
E. Condenser
F. Heat Lamp
G. Condenser Valve
H. Condenser Valve
I. Distillation Manifold

J. Condenser Valve
K. Drain Valve
L. Reaction Flasks
M. Steam Valves
N. Steam Manifold
O.T-Valve (Air/N2 /Steam)
P. Steam Reservoir
Q. Heat Lamp
R. Micro Kjeldahl Heat Mantle

Figure 1. Schwab and

Dutton deodorization
equipment modified
by Moulton.
D EODORIZATION
PRINCIPLES

The theory of edible oil

deodorization was first
described by A.E. Bailey
in 1941 and has been presented in a number of
reviews, most recently
by S.C. Loft in 1990.
Deodorization, in principle, is governed by basic
laws of distillation involving Daltons law
of partial pressure and Raoults law governing liquid and vapor phase equilibria.
Variables affecting deodorization include
temperature, time, pressure (vacuum), agitation, and stripping steam. Steam or an
inert gas is employed as a carrier to increase
agitation, break down fatty hydroperoxides, and distill volatile components from
the nonvolatile triglyceride oil.
T YPES OF DEODORIZERS
Commercial steam deodorizers can be
batch, semicontinuous, or continuous, but
batch type deodorizers are the most commonly used for laboratory work.
Continuous laboratory deodorizers were
described by Bitner et al. in 1986 and
Smouse in 1997. Batch deodorizers bubble
steam through heated oil for longer periods of time (13). In continuous deodorizers, oil passes countercurrent to the steam
as it flows through a series of distillation
trays. Residence time in a commercial continuous deodorizer varies from 20 to 30
min. Although continuous deodorizers are
more applicable than batch to current
practices, there are very few continuous
laboratory deodorizers in use today.
Hidden behind this paradox between
the processes, commercial and laboratory
batch, semicontinuous, or continuous
deodorizers is the inherent problem of
comparing data. Batch deodorizers tend to
be used when comparisons can be based on
relative numbers. Continuous deodorizers

657

Table 1.Typical conditions for batch vs. continuous

laboratory deodorizers
Process stage
Deaeration

Time
Temperature
Vacuum
Steam stripping Time
Temperature
Vacuum
Steam rate
Holding
Time
Temperature
Vacuum
Cool-down
Filter
Sparge
Freeze

Batch
<30 min
60100C
<1 mm Hg
14 hr
250275C
<1 mm Hg
310%
<1 hr
<100C
<1 mm Hg
Yes
N2
Yes

are usually employed when more nearly

absolute numbers are needed. Smouse presented many of the various designs of continuous and semicontinuous laboratory
deodorizers. The cost estimates of such
designs range between $10,000 and
$15,000. Batch deodorizers tend to be less
costly and can be fabricated from more
readily available equipment than continuous systems. The simplest design of a
batch-type laboratory deodorizer was
described by Mounts and colleagues in
1969 at an estimated price of $3,000 to
$4,000 and consisted of a high-vacuum
pump, Dewar condenser, three-necked distillation flask, thermometer, manometer,
and heating mantle.
Descriptors for laboratory deodorizers
include laboratory, micro, mini, bench, and
smallscale. It is confusing when the capacities of such equipment range from gram
quantities up to 50 kg. For clarity, we suggest that a laboratory deodorizer be defined
as one whose capacity ranges from 60 to
5000 g. As a further dividing line, we suggest that the term micro apply to deodorizers with <200-g capacity and common
apply to deodorizers with 200 to 5000-g
capacity. This separation by capacity will
differentiate between deodorizers that will
and will not have oil capacity to produce
enough oil for organoleptic evaluations.
Generally speaking, sensory testing of initial quality and after accelerated storage
requires about 300 g of oil.

T YPICAL CONDITIONS
Deodorization in simple
terms is a high-tempera<30 min
ture, high-vacuum, steam
60100C
distillation of low molec<1 mm Hg
ular weight volatiles
39 min
from nonvolatile triglyc200260C
erides. The process,
<1 mm Hg
batch or continuous, can
1.312.1%
be broken down into
<30 sec
four stages: de-aeration,
<60C
steam stripping, holding,
<1 mm Hg
and cool-down. Typical
No
reference
conditions for laboratoN2
ry deodorization are listYes
ed in Table 1.
Although
many
workers utilize silicone grease for high-vacuum work, deodorizer joints lubricated
with silicone grease can contaminate the
oil. In such circumstances, the deodorized
oil may become cloudy and exhibit elevated silicone levels. When using silicone
grease, the best practice is to use as little
grease as possible at joints coming into contact with hot oil over prolonged periods of
time or to use a nonsilicone grease such as
vegetable oil-based grease or a high molecular weight/polymer-based grease, e.g.,
Celvacene or Lubriseal (Fischer Scientific,
Itasca, IL).
The nature of the coolant used in the
condenser is important to the maintenance
and care of the vacuum pumps used in the
system. Deodorizer distillates contain
water and many products that can affect
the viscosity and performance of the vacuum pump oil. Many of the commonly used
salt and ice slurries do not have the cooling
capacity to condense the low molecular
weight contaminants. Acetone with dry ice
is one of the more common coolants used.

Continuous/
semicontinuous

C OMMON APPLICATIONS
Selection of laboratory deodorizers should
be based on four variables: degree of confidence in or reproducibility of data, type of
data required (relative compared with
absolute), capacity needed to complete the
analytical workup, and cost or availability
of equipment.
The underlying objective in any oil-pro-

cessing project is to establish a standardized

means to compare results between data sets.
Applications for laboratory deodorizers,
whether based on relative or absolute numbers, must be proven reproducible.
A majority of all deodorization work can
be performed using a batch deodorizer.
Relative effects of process changes upstream
of the deodorization stage, such as clay
dosage, clay types, water wash compared
with no water wash, etc., can be determined
using batch deodorization. Process effects of
deodorizer conditions on a given oil require
the type of absolute numbers that only a continuous deodorizer can provide.
Capacity and cost are two critical issues.
Advances in instrumentation and micro
techniques have dramatically reduced the
amount of oil needed to evaluate quality.
Today, analyses for color, chlorophyll, peroxide value (PV), free fatty acid (FFA),
oxidative stability index (OSI), fatty acid
composition by gas chromatography (GC),
and volatiles can all be performed on a
minimum of 60 g of oil.

The cost of laboratory deodorizers is

directly proportional to the capacity
needed. In most studies, batch-type
deodorizers are more commonly used
because the capacities needed fall within
that of common glassware and heating
units. Batch deodorizers also tend to cost
less than continuous deodorizers. A simple batch deodorizer setup with 100- to
500-g capacity will cost about $4000 per
unit. Single 11-L units described by
Moulton also approach $4000 per unit.
The cost of a single continuous deodorizer unit approaches $10,000 to $15,000.
One other aspect of continuous units is
that cleaning protocols can be time-consuming due to the complexity of the equipment. Batch deodorizers tend to have simpler parts that can be interchangeable and
purchased in quantities that allow for multiple instead of single runs per day.
S TATISTICAL

APPLICATIONS

Schwab and Dutton published one of the

most-cited regarding laboratory deodorizers (Figure 1), which offers a standardized
method to deodorize four independent
Volume 13 August 2002 inform

Instrumentation
658

sample (75100g) of refined bleached oil is

transferred to the reaction flask. The flask is
mounted on the heater and connected to the
steam and distillation manifolds. The sample is de-aerated for 15 min. under vacuum
(<1 mm Hg). The sample is then heated and
sparged with steam for 13 hours at temperatures of 240270C. The sample is then
cooled under vacuum. Steam is shut off
when the sample reaches 100C. Cooling is
continued until the oil temperature reaches
60C. Vacuum is then broken with nitrogen.
Each sample is then filtered through fluted
filter paper (Fisherbrand P8 Qualitative,
Table 2. Comparison of color data across each port
Coarse, Fast Flow;
of the new deodorizer
Fisher Scientific) and
RBD
nitrogen-sparged
red color a
before being sealed in
Sample # Port 1 Port 2 Port 3 Port 4 Port 5 Port 6
bottles and stored in a
1
1.0
1.0
1.1
1.1
1.0
1.1
freezer (5C).
2
1.0
0.9
0.9
0.9
1.0
1.0
Performance eval3
2.3
2.2
2.3
2.3
2.3
2.3
uations were exam4
0.8
0.9
1.0
0.9
1.0
0.8
ined in a series of tests.
RBD
b
yellow color
T EST ONE
Sample # Port 1 Port 2 Port 3 Port 4 Port 5 Port 6
Four bleached soy1
6.9
6.6
7.3
7.3
6.8
6.8
bean oil samples were
2
8.4
7.9
8.3
8.3
8.6
8.4
deodorized at each
3
14.0
13.0
13.0
13.0
14.0
13.0
port under the fol4
13.0
13.0
13.0
13.0
13.0
13.0
lowing standard set
a Analysis of variance (ANOVA) of refined, bleached, deodorized
of conditions: vacu(RBD) red color Fcal<Fcrit , = 0.5 (0.0118 < 2.773).
um, <1mm Hg; steam
b ANOVA of RBD yellow color Fcal<Fcrit , = 0.5 (0.0077 < 2.773).
rate, 5%; 260C; and
residence time, 1
The basic design of the microdeodorizer hour. Analysis of variance of the Lovibond
consists of the following: steam generator, red and yellow color val-ues (AOCS
steam manifold, reaction flask, six-port Official Method Cc 13b-45, Automated
heater, distillation manifold, condenser, Lovibond Tintometer PFX990) showed
vacuum gauge, and high-vacuum pump that there was no difference between oils
(Fig. 2, Diagrams 1and 2).
generated in each port (Table 2).
The microdeodorizer was assembled for
$6,500. Fifty percent of the cost went for T EST TWO
customizing the glassware (Universal Refined canola oil was treated/bleached
Instruments, Palatine, Illinois). The remain- under multiple conditions followed by subder of the cost was distributed between the sequent deodorization in duplicates. Port
Micro Kjeldahl Heater (Labconco, Kansas positions and selection of samples were
City, MO), the Welch Dual Seal model randomized. Conditions were the same as
1400 vacuum pump (Fisher Scientific), and Test 1. Pooled data from 95 pairs of data
miscellaneous supplies (e.g., thermometers, sets resulted in a standard deviation of
highvacuum tubing, clamps, etc.).
0.09 for red colors (range 0.4 to 3.5 ) and
The following protocol can be used. A 1.83 for yellow colors (range 2.3 to 70).
samples simultaneously under identical
conditions of time, temperature, and steam
stripping.
Currently, with the evolution of statistics and multiple-variable experimental
designs, there is a demand to deodorize
more samples in shorter times. To meet this
demand, we introduce a modified microsized version of the Schwab deodorizer
(Figure 2). Advantages of this design are
that it provides simultaneous deodorization
of six 7510-g, independent samples in
three to four hours.

Volume 13 August 2002 inform

T EST THREE
Standardization in the case of both the
microdeodorizer and the Schwab deodorizer applies more to the step-by-step protocol
than to creation of absolute values. In practice these deodorizers are used more for
generating relative values than absolute
ones. This is evident when industrial and
laboratory deodorized oil results under a
standard set of conditions are compared.
Matched sets of bleached and deodorized
oil samples from various producers were
deodorized in the laboratory under the
same conditions as Test 1. Color, FFA, and
PV analyses using Official AOCS Methods
are reported in Table 3.
It would be desirable to standardize
deodorization conditions to be applicable
to all oils. However, variations in feed
stock quality, industrial equipment, and
process conditions as well as differences
due to the nature of each given oil make it
impractical to do so. A comparison of laboratory and plant data, however, shows
good correlation between the two data sets.
Closer scrutiny of the data shows that the
base range is probably too broad.
Inspection of each individual oil group
shows that, as expected, some diversity
exists between laboratory and plant data.
In general, the plant data had higher PV
and FFA values than the laboratory data.
This can be attributed to the fact that the
deodorized oil samples from the plants
were analyzed after 2 to 20 days after shipping and handling. Commercial oil will
also contain up to 50 ppm citric acid,
which will affect the final FFA content.
Corn oil showed the most variation
between laboratory and plant data. In circumstances such as these, one would have
to set up an experimental matrix to establish the best set of conditions to attempt to
make absolute comparisons between plant
and laboratory data.
A simple test was conducted to demonstrate the diversity of treating the bleached
corn oil samples from the different producers. Wet-milled corn oil samples from producers I and K and dry-milled corn oil from
producer L were subjected to a simple experiment where deodorization time and tem-

659

Table 3. Data on deodorization in a plant vs. laboratory

RBD data
Oil
Type

Producer Deod
Temp
(C)

Palm

A
A
A
B
B
B
C
Soybean D
E
F
E
G
Canola
H
H
H
H
H
H
H
Peanut
D
Rice
D
Corn
I
I
I
I
K
L
Deodorization

Red
Lab
scale

260 4.4
260 3.3
260 3
260 2.9
260 2.9
260 3.25
260 2.9
260 0.85
260 0.4
260 0.4
260 0.4
260 1.1
260 0.8
260 0.85
260 0.75
260 0.9
260 0.75
260 0.75
260 0.9
260 0.6
260 3.45
260 1.1
260 1.2
260 1.2
260 1.2
260 2.1
260 3.5
time, vacuum,

Yellow

Plant
scale

Lab
scale

Plant
scale

% FFA

Chlorophyll

4.5 0.1
70
70
0
3.1
0.2
50
50
0
3.2 0.2
50
38
12
2.3
0.6
43.5
24
19.5
2.4
0.5
42.5
28
14.5
2.4
0.85 50
24
26
3
0.1
50
50
0
0.8
0.05
6.3
4.7
1.6
0.6 0.2
2.7
5.2 2.5
0.5 0.1
2.35
5.4 3.05
0.4
0
2.7
2.7
0
0.8
0.3
5.5
5.3
0.2
0.7
0.1
5.6
5.4
0.2
0.7
0.15
6.55
5.3
1.25
0.7
0.05
5.4
5.5 0.1
0.8
0.1
6.8
5.7
1.1
0.8 0.05
5
5.3 0.3
0.9 0.15
5.6
6.4 0.8
1
0.1
6.8
9.1 2.3
1
0.4
3.2
4.55 1.35
3.2
0.25 30
32
2
1.5 0.4
6.9
9.9 3
1.4 0.2
7.6
9.7 2.1
1.6 0.4
7.6
12
4.4
1.7 0.5
8.8
12
3.2
2.5 0.4
15
26
11
2.4
1.1
26
17
9
and temperature the same: 1 hr at ~1.0

0
0
0
0
0
0
0
26
21
16
28
34
19
21
25
24
11
19
24
0
0
0
0
0
0
0
0
mm Hg

Lab
scale

Plant
scale

0.038
0.120
0.018
0.080
0.032
0.100
0.026
0.100
0.080
0.100
0.100
0.120
0.000
0.007
0.023
0.023
0.025
0.025
0.028
0.031
0.011
0.010
0.010
0.023
0.010
0.010
0.021
0.015
0.018
0.020
0.010
0.010
0.005
0.010
0.010
0.005
0.031
0.020
0.310
0.440
0.064
0.038
0.331
0.054
0.441
0.059
0.345
0.050
0.204
0.056
0.113
0.073
0.298
0.169
and 260C

PV

Lab
scale

Plant
scale

0.082
0.062
0.068
0.074
0.020
0.020
0.007
0.000
0.000
0.003
0.001
0.013
0.000
0.006
0.002
0.000
0.005
0.005
0.011
0.130
0.026
0.277
0.382
0.295
0.148
0.040
0.129

0
0
0
0.05
0
0
0
0
0
0
0.05
0

0.2
0.1
0.05
0.15
0.15
0.1
0.1
0.05
0.05
0
0
0
0
0.05
0
0
0.05
0
0.05
0
0
0.05
0.05
0
0.05
0.21
0

0.2
0.1
0.05
0.1
0.15
0.1
0.1
0.05
0.05
0
0.05
0
0
0.05
0
0
0.05
0
0
0
0
0.05
0
0
0.05
0.21
0

0
0
0
0
0
0.05
0
0
0
0.05
0
0
0
0

Table 4. Comparison of RBD corn oil data

Producer

Process

Conditions

FFA

OSI

Color red/yellow

Deodorizer

1 hr

2 hr

1 hr

2 hr

1 hr

2 hr

250
260
270

0.441
0.062
0.059

0.113
0.078
0.07

0.298
0.331
0.169

3.00
5.75
6.10

5.30
5.45
4.25

3.45
2.93
3.95

1.2 / 7.6
1.2 / 6.0
1.4 / 9.7

2.1 / 15
2.0 / 12
2.5 / 26

3.5 / 26
3.4 / 26
2.4 / 17

Wet

Laboratory

Dry

Plant
Laboratory

Wet

Plant
Laboratory
Plant

250
260
270
250
260
270

0.122

1.63

0.160

5.18

1.73

3.43

2.1 / 13

3.5 / 28

2.9 / 21

Volume 13 August 2002 inform

Instrumentation
660

perature were varied. The results are shown

in Table 4. Color and FFA levels agreed with
the values from the plant sample after one
hour at 260C. The wet-milled samples
required higher temperatures and longer
times to compare with the plant samples.
The practical value of the microdeodorizer is apparent when comparing relative
numbers. For example, Table 5 lists the relative data from a plant trial in which various bleaching clays were used to treat
canola oils. The OSI of the RBD oils from
the plant were compared to the data from
laboratory deodorized samples using the
bleached oil samples from the plant.
Although there is a definite difference
between the absolute numbers, the relative
order of stability was maintained from
plant to laboratory.
REFERENCES
Bailey, A.E., Steam Deodorization of Edible
Fats and Oils, Ind. Eng. Chem. 33:
404408 (1941).
Loft, S.C., DeodorizationTheory and
Practice, in World Conference
Proceedings Edible Fats and Oils

Processing: Basic Principles

Table 5
and Modern Practices, editRank data of laboratory vs. plant OSI data
ed by D.R. Erickson,
Sample Laboratory Laboratory Average Rank order
American Oil Chemists
OSI 1
OSI 2
OSI
(1 = best)
Society, Champaign, Illinois,
1
7.85
7.60
7.73
3
pp. 117 123, 1990.
2
6.70
7.25
6.98
4
Mounts, T.L., C.D. Evans, H.J.
3
8.15
8.45
8.3
2
Dutton, and J.C. Cowan,
4
8.55
8.55
8.55
1
Some Radiochemical Experi5
6.10
6.85
6.3
5
ments on Minor ConstiSample
Plant
Plant
Average Rank order
tuents in Soybean Oil, J.
OSI 1
OSI 2
OSI
(1 = best)
Am. Oil Chem. Soc.
1
8.10
7.95
8.03
3
46:482484 (1969).
2
7.50
7.15
7.33
4
Schwab, A.W., and H.J.
3
8.55
8.75
8.65
2
4
8.95
8.85
8.90
1
Dutton,
The
Flavor
5
7.30
6.40
6.85
5
Problem of Soybean Oil. III.
A Four-Sample, Glass
Deodorizer for Vegetable Oils, J. Am.
Laboratory Deodorizer, J. Am. Oil
Oil Chem. Soc. 63:338340 (1986).
Chem. Soc. 25: 5759 (1948).
Moulton,
K.J.
Sr.,
Laboratory Smouse, T.H., A Laboratory Continuous
Deodorizer, inform 8:11761181
Deodorization of Vegetable Oil, J. Am.
(1997).
Oil Chem. Soc,. 66:302308 (1989).
Shukla, V.K.S., Description of the IFSC
minideodorizer, inform 8:11821183 Readers may reach the authors by contacting
(1997).
David D. Brooks at Oil Dri Corp. of America,
Bitner, E.D., J.P. Friedrich, and T.L. 777 Forest Edge Dr., Vernon Hills, IL 60061Mounts, Laboratory Continuous 3197, e-mail: [email protected]

Advertisement

Volume 13 August 2002 inform