Pilot plant fermentation of high test molasses compared with blackstrap molasses using factorial design


Cacho, E.; Murphy, N.F.; Fontanet, E.

Journal of Agriculture of the University of Puerto Rico 72(1): 9-17

1988


In a 2-level Factorial Design study for direct statistical comparison of the performance of high test molasses (HTM) and blackstrap molasses (BM) as fermentation substrates, the dependent variables were alcohol yield and congener formation for 24-h fermentation. Another purpose was to identify those independent variables (and the interactions of those variables) which have significant effects over the dependent variables. The independent variables were fermentation substrate, initial yeast cell count, molasses feed addition time and quantity of molasses feed added. Sixteen pilot plant scale (3785 litre) fed batch fermentations were performed. Variables which showed a statistically significant (P = 0.05) effect on alcohol yield and congener generation were fermentation substrate and quantity of molasses feed added. A significant effect of fermentation substrate on pH and total acidity was also observed. In general BM fermentations were in a more advanced stage and produced higher fusel content than HTM fermentations after 24 h.

Pilot
plant
fermentation
of
high
test
molasses
compared
with
blackstrap
molasses
using
factorial
design'
Eduardo
Cacho,
Nivia
F.
Murphy
and
Eleanor
Fontanet
2
ASTRACT
A
two-level
Factorial
Design
study
was
done
for
the
purpose
of
making
a
direct
statistical
comparison
of
the
performance
of
high
test
molasses
(HTM)
and
blackstrap
molasses
(BM)
as
fermentation
substrates.
The
de-
pendent
variables
under
study
were
alcohol
yield
and
congener
formation
for
24-hour
fermentation.
Another
purpose
was
to
identify
those
independ-
ent
variables
(and
the
interactions
of
those
variables)
which
have
signifi-
cant
effects
over
the
dependent
variables
under
study.
The
independent
variables
were
fermentation
substrate,
initial
yeast
cell
count,
molasses
feed
addition
time
and
quantity
of
molasses
feed
added.
Sixteen
pilot
plant
scale
(3785
L)
fed
batch
fermentations
were
performed.
Variables
which
showed
a
statistically
significant
(p
=
0.05)
effect
on
alcohol
yield
and
congener
generation
were
fermentation
substrate
and
quantity
of
molasses
feed
added.
A
significant
effect
of
fermentation
substrate
on
pH
and
total
acidity
was
also
observed.
In
general
BM
fermentations
were
in
a
more
advanced
stage
and
produced
higher
fusel
content
than
HTM
fer-
mentations
at
24
hr
of
fermentation.
INTRODUCTION
The
rum
industry
of
Puerto
Rico
has
consistently
accounted
for
about
10%
of
the
gross
insular
income,
making
it
the
second
source
of
revenue
of
the
island.
Traditionally
the
Puerto
Rican
rum
industry
has
used
blackstrap
molasses
(BM),
a
by-product
of
the
sugar
industry,
as
raw
material
because
of
its
availability
and
price.
However,
since
1970
the
local
sugar
industry
has
not
been
able
to
supply
all
the
BM
needed
by
the
rum
industry.
In
1983
rum
producers
imported
36,749,724
gal.
(1,390,977
hL)
of
BM,
94%
of
the
total
used.
BM
is
imported
from
other
countries,
some
of
which
are
at
present
increasing
their
rum
production
to
take
advantage
of
the
recently
relaxed
U.
S.
tariff
barriers.
This
situation
has
caused
great
concern
among
Puerto
Rican
rum
producers.
'Manuscript
submitted
to
Editorial
Board
20
August
1986.
2
Chemical
engineer,
bacteriologist,
Rum
Pilot
Plant,
and
assistant
statistician,
Statistics
Section,
Agricultural
Experiment
Station,
MayagUez
Campus,
University
of
Puerto
Rico,
Rio
Piedras,
P.
R.
The
authors
express
their
deepest
appreciation
to
Mr.
Amador
Belardo,
Technical
Director,
Rum
Pilot
Plant,
Agricultural
Experiment
Station,
University
of
Puerto
Rico,
for
invaluable
advice,
and
to
Dr.
Guillermo
Martinez,
Zulma
de
Ayala,
Marta
Rivera
and
Luis
Garcia
of
the
Chemistry
Section
of
the
Rum
Pilot
Plant
for
their
fine
help
in
the
analytical
work
required
in
these
studies.
9
10
CACHO
ET
AL./FERMENTATION
OF
HIGH
TEST
MOLASSES
The
Rum
Pilot
Plant
has
been
working
since
1979
on
possible
alterna-
tives
to
BM
as
raw
material
which
could
be
produced
locally
to
eliminate
dependency
on
outside
sources.
Emphasis
has
been
placed
on
methods
for
the
production
and
evaluation
of
high
test
molasses
(HTM).
HTM
is
defined
as
a
clarified
sugarcane
syrup,
partially
inverted
to
avoid
crys-
talization,
and
evaporated
to
ca.
85°
Brix.
A
procedure
for
manufacturing
HTM
was
developed
at
the
Rum
Pilot
Plant
(15),
and
over
the
past
5
years
comparative
studies
of
HTM
and
BM
have
concentrated
on
nutri-
ents,
rate
of
fermentation,
congener
generation,
characterization
of
slops
and
aging
of
distillates
(8,10,13,18).
These
studies
have
demonstrated
that
there
are
significant
differences
between
HTM
and
BM
in
many
aspects
of
rum
manufacture.
For
this
study
a
set
of
experiments
was
designed
that
allowed
direct
statistical
comparison
of
the
performance
of
HTM
and
BM
as
fermenta-
tion
substrates
on
a
pilot
plant
scale
while
concurrently
screening
for
the
effects
and
the
interactions
of
other
independent
variables
studied.
The
dependent
variables
under
study
were
alcohol
yield
and
congener
gener-
ation.
Obviously,
the
industry
would
benefit
from
faster,
more
efficient
fermentation
but
in
terms
of
an
economic
analysis
the
quality
of
the
product
also
has
to
be
taken
into
account.
Puerto
Rican
light
rums
have
traditionally
been
characterized
by
lower
fusel
content
(propyl,
isop-
ropyl,
n-butyl,
isobutyl,
amyl,
isoamyl,
d-amyl,
and
other
high
boilers)
than
that
of
heavy
type
rums,
brandies
and
whiskeys
(2).
Hence,
accord-
ing
to
Brau,
decreasing
the
amount
of
fusel
oil
produced
during
the
fer-
mentation
of
process
to
a
minimum
would
be
most
desirable
(3).
Brau
(4)
TABLE
1.—Factorial
design
experimental
runs
Runs
X,
X
2
X
s
X
4
Raw
material
Cell
count
Addition
time
Gallons
added
F.D.
1
HTM
High
High
High
F.D.
2
HTM
High
High
Low
F.D.
3
HTM
High
Low
High
F.D.
4
HTM
High
Low
Low
F.
D.
5
HTM
Low
High
High
F.D.
6
HTM
Low
High
Low
F.D.
7
HTM
Low
Low
High
F.D.
8
HTM
Low
Low
Low
F.D.
9
BM
High
High
High
F.D.
10
BM
High
High
Low
F.D.
11
BM
High
Low
High
F.D.
12
BM
High
Low
Low
F.D.
13
BM
Low
High
High
F.D.
14
BM
Low
High
Low
F.D.
15
BM
Low
Low
High
F.D.
16
BM
Low
Low
Low
J.
Agric.
Univ.
P.R.
VOL.
72,
NO.
1,
JANUARY,
1988
11
also
pointed
to
the
importance
of
the
recovery
of
desirable
congeneric
substances,
such
as
esters
and
acids.
Martinez
(8)
in
his
laboratory
scale
studies
found
that
HTM
resulted
in
products
higher
in
esters
and
lower
in
fusel.
We
desired
to
determine
whether
these
favorable
results
could
be
replicated
on
a
pilot
plant
scale.
MATERIALS
AND
METHODS
Experimental
design
A
two-level
factorial
design
consists
of
high
and
low
levels
of
each
independent
variable
in
all
combinations
(table
1).
For
N
variables
this
results
in
2
N
runs.
For
each
variable
a
measure
of
the
error
variance
and
effect
variance
is
made.
If
there
is
no
real
effect
and
the
calculated
vari-
ance
is
merely
a
measure
of
error,
large
differences
(as
determined
by
F-distribution)
between
the
effect
variance
and
the
error
variance
are
improbable
(6).
We
assigned
the
run
order
randomly
to
assure
the
valid-
ity
of
the
results
and
performed
a
run
test
(17)
to
corroborate
that
the
order
was
indeed
random
(table
2).
The
two-level
factorial
design
was
used
to
study
the
effects
of
the
independent
variables
(raw
materials
(HTM
&
BM),
initial
yeast
cell
count,
molasses
feed
addition
time,
and
quantity
of
molasses
feed
added)
on
the
dependent
variables
of
interest
(alcohol
yield,
residual
sugars
as
invert,
pH,
total
acidity
as
acetic,
and
other
congeners
generated)
all
of
these
at
24-hour
fermentation
time.
TABLE
2.—Run-test—Test
for
randomness
of
runs
as
performed
Runs
as
per
factorial
Runs
as
performed
X,
X
2
X
3
x4
design
A
+'
F.D.
4
B
+ +
+
F.D.
9
C
+
+
+
F.D.
5
D
+
+
F.D.
11
E
+
F.D.
14
F
+
+
F.D.
13
G
+
F.D.
15
H
+
F.D.
8
I
+
+
+
+
F.D.
1
J
+
+
+
F.D.
2
K
+
F.D.
12
L
+
+
F.D.
10
M
+
+
F.D.
7
N
-,-
+
+
F.D.
3
O
F.D.
16
P
+
+
F.D.
6
Runs
(r)
9
8
10
9
1
+
indicates
high
level
of
the
variab
le
and
indicates
low
level
of
the
variable.
12
CACHO
ET
AL./FERMENTATION
OF
HIGH
TEST
MOLASSES
Factorial
design
variables
and
levels
X
1
=
Raw
material
Levels
High
=
High
test
molasses
(HTM)
Low
=
Blackstrap
molasses
(BM)
X2
=
Initial
yeast
cell
count
Levels
High
=
200
gallons
(757
L)
with
270
x
10
6
cell/ml
of
Sac-
charomyces
cerevisiae
yeast
strain
PPR-80
Low
=
100
gallons
(379
L)
with
270
x
10
6
cell/ml
of
Sac-
charomyces
cerevisiae
yeast
strain
PPR-80
X3
=
Molasses
feed
addition
time
Levels
High
=
total
quantity
added
in
12
hours
at
constant
rate
Low
=
total
quantity
added
in
16
hours
at
constant
rate
X4
=
Quantity
of
molasses
feed
added
Levels
High
=
290
gallons
(1097
L)
of
approximately
50
g/100
ml
total
sugars
as
invert;
theoretically
should
yield
10%
v/v
al-
coholic
solution
Low
=
230
gallons
(871
L)
of
approximately
50
g/100
ml
total
sugars
as
invert;
theoretically
should
yield
8%
v/v
al-
coholic
solution.
Equipment
Fermentations
were
carried
out
in
3785
L
(working
capacity)
tanks
with
covered
tops
to
achieve
anaerobic
conditions.
These
tanks
were
equipped
with
a
2-inch
diameter
outlet
for
the
CO
2
generated,
a
65
ft
2
surface
area
coil
for
cooling
water,
a
sight
gage
for
measuring
liquid
level,
a
thermometer,
and
1.5-hp
centrifugal
pump
for
recirculating
the
fermenting
broth.
We
added
the
molasses
feed
at
a
constant
rate
by
pumping
from
a
1900
L
(working
capacity)
tank
with
sanitary
metering
pump
equipped
with
a
3/4
hp
variable
speed
drive.
Flow
rate
was
ad-
justed
manually
and
measured
with
a
rotameter.
The
molasses
mash
and
fermented
broth
were
centrifuged
with
a
Westfalia
Separator
Model
NA-
7-06-076.
3
Distillation
was
performed
with
the
Rum
Pilot
Plant
Beer
Col-
umn.
Aside
from
the
distillation
column
all
other
materials
of
construc-
tion
in
direct
contact
with
the
molasses
feed
or
broth
were
of
type
316
stainless
steel.
'Trade
names
in
this
publication
are
used
only
to
provide
specific
information.
Mention
of
a
trade
name
does
not
constitute
a
warranty
of
equipment
of
materials
by
the
Agricul-
tural
Experiment
Station
of
the
University
of
Puerto
Rico,
nor
is
this
mention
a
statement
of
preference
over
other
equipment
or
materials.
J.
Agric.
Univ.
P.R.
VOL.
72,
NO.
1,
JANUARY,
1988
13
Experimental
procedures
Initially
all
the
equipment
was
sterilized
with
live
steam.
The
fermen-
tation
feed
consisted
of
a
50
g/100
ml
fermentable
sugar
solution
(with
HTM
or
BM
accordingly).
Ammonium
sulfate,
1.5
g/L
for
BM
and
2
g/L
for
HTM,
was
added
as
nitrogen
source;
pH
was
adjusted
with
sulfuric
acid
to
4.7.
4
The
molasses
feed
was
pasteurized
at
77°
C
for
45
min
and
then
cooled
to
30°
C
with
cooling
water
through
the
internal
coil
heat
exchanger.
The
feed
was
then
centrifuged
and
ready
for
use
in
the
fer-
mentation.
We
prepared
the
yeast
inoculum
by
scaling
up
the
procedures
used
in
the
laboratory
scale
studies
(10).
Yeast
cell
count
was
determined
on
a
Newbauer
Hemacytometer.
The
fermentor
was
first
charged
with
the
amount
of
tap
water
re-
quired
for
each
experiment;
subsequently,
the
yeast
inoculum
was
added,
followed
by
the
molasses
feed,
which
was
added
at
a
constant
rate.
The
final
volume
for
all
the
runs
was
3785
L.
All
the
experiments
ran
for
24
hours
starting
from
the
molasses
feed
addition.
Throughout
the
experiment
the
recirculating
pump
was
running
and
the
temperature
in
the
fermenter
was
kept
at
86
±
F
(30°
C)
with
cooling
water
through
the
inner
coil.
To
monitor
the
process,
we
drew
samples
every
4
hours
and
analyzed
them
for
total
sugars
as
invert
(g/100
ml),
%
alcohol
(v/v),
total
acidity
(g/L),
and
pH
as
described
in
the
Official
Analytical
Methods
of
the
Rum
Pilot
Plant
(12).
In
addition,
the
24-hour
samples
were
analyzed
for
other
congeners
by
gas
chromatography
(8).
To
stop
the
fermentation,
we
added
2
g/L
of
mercuric
chloride
to
the
samples.
At
the
end
of
24
hours
the
fermentation
broth
was
centrifuged
and
then
distilled.
In
order
to
keep
interfering
variables
to
a
minimum,
we
based
the
statistical
studies
on
the
24-hour
sample
before
centrifugation
and
distillation.
Statistical
analysis
The
data
for
the
24-hour
sample
were
analyzed
with
a
System
34
IBM
computer.
Since
no
replications
were
done,
the
error
estimate
used
higher
order
interactions
to
estimate
experimental
error
as
follows
(5):
the
four
three-factor
interactions
and
the
four-factor
interaction
were
pooled
to
estimate
the
experimental
error.
RESULTS
AND
DISCUSSION
By
means
of
the
Run-test
it
was
determined
there
was
randomness
of
assignment
in
terms
of
high
and
low
treatments
for
each
of
the
vari-
ables,
as
shown
in
table
2.
The
critical
values
of
(r)
show
that
we
should
accept
the
null
hypothesis
of
randomness
(p=
.05)
if
4
<
r
<
14.
There-
fore,
the
null
hypothesis
of
randomness
cannot
be
rejected.
'Optimum,
as
reported
by
Murphy
(10).
rP
C
ACHO
ET
AL
./
FERMENTATI
ON
OF
HIGH
TEST
MOL
AS
SES
TABLE
3.-Characterization
of
24-hour
fermented
broth
Run
%
of
theore-
tical
yield
pH
Acidity
g/1
Alcohol
%
(v/v)
Total
sugar
as
invert
000
ml
Congeners
mg/100
ml
at
80°P
Acetal-
dehyde
Methyl
acetate
Ethyl
acetate
Acetal
Propyl
alcohol
Isobutyl
alcohol
Isoamyl
acetate
1-Butyl
alcohol
Isoamyl
alcohol
Amyl
alcohol
F.D.
1'
71.96
3.67
4.93
8.7
4.33
15.4
10.6
5.8
0
32.5
10.8
2.9
1.2
27.0
0.4
F.D.
2
80.48
3.59
4.86
7.8
2.26
10.6
9.6
6.5
0
25.8
8.1
4.5
2.5
28.8
0.5
F.D.
3
71.64
3.57
4.94
8.5
4.11
10.8
5.6
0
2.3
23.0
8.4
0
6.5
16.0
0
F.D.
4
90.84
3.61
4.36
7.8
1.12
14.9
8
8.4
8
4.9
3
1.0
3
23.5
3
9.2
3
1.5
3
3.5
8
24.8
8
0.3
3
F.D.
5
75.00
3.86
4.73
7.3
4.65
18.6
5.5
7.7
0
26.3
4.9
1.1
6.0
24.1
0.5
F.D.
6
72.65
3.80
4.04
6.1
2.90
12.2
5.1
2.5
0
14.8
7.7
0
0.6
23.2
0
F.D.
7
72.45
3.51
4.64
7.1
4.77
8.8
8.6
0
4.6
15.0
11.4
0
6.2
22.8
0
F.D.
8
88.47
3.61
4.43
6.8
1.95
28.1
13.5
11.8
0
27.0
12.9
2.2
1.6
31.5.
0.5
F.D.
9
2
92.63
4.35
7.04
10.6
1.35
4.5
4.5
3.4
0
22.9
15.6
0
1.1
48.4
0.8
F.D.
10
89.03
4.16
5.88
8.2
1.01
7.8
4.8
0
0
29.2
13.6
0 0
40.4
0
F.D.
11
93.13
4.13
7.08
9.7
1.28
4.9
2.9
3.7
0
22.3
12.4
0.4
3.3
44.5
0.4
F.D.
12
93.01
4.11
4.84
8.1
1.20
5.9
3.9
0
0
23.7
14.8
0
0
40.4
0
F.D.
13
94.01
4.18
6.39
8.9
2.45
9.4
11.6
11.6
0
27.5
20.2
3.4
0.4
57.6
1.2
F.D.
14
96.11
4.19
5.56
8.2
1.10
4.9
2.9
2.9
0
19.0
15.1
0
1.0
45.4
0.5
F.D.
15
90.81
4.37
5.98
8.7
2.80
7.5
10.2
6.6
0
21.7
24.8
2.6
1.3
80.8
1.3
F.D.
16
94.71
4.53
5.80
7.5
1.05
9.3
6.5
0
0
19.7
16.9
0
0
37.8
16.4
Averages
and
standard
deviation
'
F.D.
1
to
F.D.
8
R=77.94
X=3.65
X=4.62
X=7.51
X=3.26
X
=14.93
X=8.36
X=4.9
X
=
.99
X=23.49
X=9.17
1=1.53
X=3.51
X=24.77
=.27
(HTM)
a=7.80
cr=0.12
a=.318
a
=
.865
a
=1.39
a
=
6.68
a
=
3.15
a
=
4.33
a
=
1.81
a=6.514
v
=2.698
a=1.75
a
=
2.61
a=5.00
a=.256
F.D.
1
to
F.D.
16
X
=
92.93
X=4.25
X
=
6.07
X=8.75
X=1.53
X=7.29
X=5.91
X
=3.53
X
=0
X=23.25
X=16.68
=0.8
X
=
.89
=49.41
X
=2.58
(BM)
a=2.20
a=.148
a
=.752
a
=
1.01
a
=.692
a
=3.01
a=3.31
a=4.01
a=0
a=3.541
a=4.025
a
=
1.38
a=1.11
a=14.10
a=5.61
3
Estimates
values.
J.
Agric.
Univ.
P.R.
VOL.
72,
NO.
1,
JANUARY,
1988
15
Table
3
shows
the
values
for
the
dependent
variables
under
study,
alcohol
yield
and
congener
generation,
other
dependent
variables
that
were
deemed
relevant
for
the
interpretation
of
the
results,
and
averages
and
standard
deviations
for
HTM
runs
and
BM
runs.
Runs
F.D.
1
to
F.D.
8
include
the
HTM
runs
and
runs
F.D.
9
to
F.D.
16
include
the
BM
runs.
For
BM
runs,
alcoholic
yield
was
higher
on
the
average
(92.93
vs.
77.94%
for
HTM)
while
the
standard
deviation
values
were
lower
(2.2
vs.
7.80
for
HTM).
Basically
for
BM
runs
the
fermentation
was
close
to
completion
for
all
runs,
whereas
for
the
HTM
runs
the
fermentation
was
in
different
stages
of
completion
at
24
hours.
Through
statistical
analysis
(table
4)
it
was
determined
that
this
effect
of
raw
material
on
yield
was
significant
(p
=
0.01).
The
average
resultant
pH
was
lower
for
the
HTM
runs
(3.65
vs.
4.25
for
BM)
even
when
the
average
total
acidity
was
lower
for
HTM
runs
(4.62
vs.
6.07
for
BM).
Through
statistical
analysis
it
was
determined
that
this
effect
of
raw
material
on
pH
and
total
acidity
was
significant
(p=0.01).
This
phenomenon
had
been
observed
by
Murphy
in
laboratory
scale
experiments
(9)
and
could
be
explained
by
the
lower
buffering
ca-
pacity
inherent
in
the
HTM.
This
lower
pH
should
be
expected
to
have
a
detrimental
effect
on
alcohol
yield
since
the
optimum
pH
for
ethanol
production
is
from
4.5
to
5.0
(1).
The
tabulation
below
confirms
the
previously
observed
trend
of
con-
gener
formation
(8)
whereby
HTM
fermentations
generate
more
esters
and
less
fusel.
HTM
BM
Acetaldehyde
14.9
7.3
Esters
14.8
10.2
Acetal
1.0
0.0
Fusel
61.2
92.8
Total
91.9
110.3
Those
congeners
that
exhibit
statistically
significant
differences
depend-
ent
on
raw
material
are
acetaldehyde,
isobutyl
alcohol,
n-butyl
alcohol
and
isoamyl
alcohol.
No
significant
difference
was
found
for
esters
(table
TABLE
4.—F
values
for
significant
effects
Independent
variables
Dependent
variables
Acetal-
Isobutyl
1-Butyl
Isoamyl
Yield
pH
Acidity
dehyde
alcohol
alcohol
alcohol
X
1
(raw
material)
51.19**'
68.44**
44.93**
8.17*
2
25.92**
8.63*
26.14**
X4
(gallons
added)
6.78*
11.78*
(P
=
0.01).
(P
=
0.05).
16
CACHO
ET
AL./FERMENTATION
OF
HIGH
TEST
MOLASSES
4).
Acetaldehyde
differences
can
be
explained
in
terms
of
a
deficiency
of
growth
substances
(11).
n-Butyl
alcohol
is
the
only
fusel
component
which
is
found
in
signficantly
higher
concentrations
in
HTM
fermentations
(av-
erage
3.51
vs.
0.89
for
BM).
Most
of
the
fusel
content
in
BM
fermenta-
tions
can
be
accounted
for
by
isoamyl
alcohol,
which
is
in
a
concentration
twice
as
high
as
in
HTM
fermentations
and
accounts
for
53.2%
of
the
fusel
in
BM
fermentations.
Table
4
shows
that
the
independent
variables
Raw
Material
(X
1
)
and
gallons
of
molasses
feed
added
(X
4
)
are
responsible
for
the
effects
regis-
tered.
Although
the
negative
effect
of
using
higher
levels
of
sugar
con-
centrations
on
yield
is
well
documented,
it
was
desired
to
determine
whether
there
were
any
interactions
with
other
independent
variables;
none
were
observed.
The
effect
of
X4
on
acidity
can
be
explained
by
the
amount
of
sulfuric
acid
needed
to
adjust
the
pH
to
4.7.
Although
the
HTM
fermentations
are
definitely
slower
under
the
con-
ditions
studied
it
is
important
to
remember
that
BM
fermentations
have
been
optimized
by
many
years
of
experimentation
and
experience,
BM
being
the
traditional
raw
material
for
rum
manufacture.
This
study
points
to
the
need
of
future
studies
to
optimize
HTM
fermentations
in
terms
of
rate
of
fermentation.
Specifically,
studies
on
nutrients
and
pH
levels
which
have
proven
important
for
other
raw
materials
(7,14)
are
necessary
for
HTM.
Also,
the
effect
of
different
levels
of
invert
sugars
(50-65%
for
HTM;
20%
for
BM)
could
be
responsible
for
different
rates
of
fermentations
(16).
They
should
be
studied.
RESFMEN
ComparaciOn
de
la
fermentaciOn
a
escala
piloto
de
mieles
ricas
y
finales
con
un
diseno
factorial
Se
neva
a
cabo
un
estudio
estadistico
tipo
diserio
factorial
a
dos
niveles
para
comparar
estadisticamente
el
comportamiento
de
dos
substratos
de
fermentaciOn:
mieles
ricas
y
mieles
finales.
Ademas
se
perseguia
identificar
las
variables
independientes
(y
las
interacciones
de
esas
variables)
que
tienen
efectos
significativos
sobre
las
variables
dependientes
que
se
men-
cionan
mos
adelante;
rendimiento
de
alcohol
y
generaciOn
de
congenericos
luego
de
24
horas
de
fermentation.
Las
variables
independientes
fueron
substrato
de
fermentaciOn,
conteo
initial
de
levadura,
tiempo
de
adiciOn
de
miel
y
cantidad
de
miel
anadida.
Para
este
propOsito
se
efectuaron
16
fermentaciones
de
tipo
semicontinuo
a
escala
de
planta
piloto
(3785
L).
Las
variables
que
tuvieron
un
efecto
estadisticamente
significativo
(p
=
0.05)
sobre
el
rendimiento
alcoholic°
y
la
generaciOn
de
congenericos
fueron
el
substrato
de
fermentation
y
la
cantidad
de
miel
anadida.
Tambien
se
ob-
serve)
un
efecto
significativo
del
substrato
de
fermentation
sobre
el
pH
y
la
acidez
total.
En
general,
las
fermentaciones
con
mieles
finales
estaban
mas
adelantadas
y
tenian
un
contenido
de
fusel
mas
alto
que
las
fermentaciones
con
miel
rica
despues
de
24
horas
de
fermentation.
J.
Agric.
Univ.
P.R.
VOL.
72,
NO.
1,
JANUARY,
1988
17
LITERATURE
CITED
1.
Anonymous,
Ethanol
fuels
reference
guide,
Solar
Energy
Research
Institute,
Golden,
Colorado.
2.
Batiz,
H.
et
al.,
1975.
Composition
de
los
roves
comerciales
de
Puerto
Rico
de
1963,
1965,
1967,
1969
y
1971,
Progr.
Rep.
1-75,
Rum
Pilot
Plant,
Agric.
Exp.
Stn.
Univ.
P.
R.,
Rio
Piedras,
P.
R.
3.
Brau,
H.,
1957.
Review
on
the
origin
and
composition
of
fusel
oil,
Technical
Paper
No.
17,
Agric.
Exp.
Stn.
Univ.
P.
R.,
Rio
Piedras,
P.
R.
4.
Brau,
H.,
1958.
Considerations
on
the
convenience
of
distilling
at
high
proof
in
continu-
ous
alcoholic
distillation
for
beverages,
J.
Agric.
Univ.
of
P.
R.
42
(4):
248-57.
5.
Cochran
and
Cox,
1984.
Experimental
Designs.
2nd
ed,
p.
219.
6.
Davis,
0.
L.,
Ed,
1954.
The
Design
and
Analysis
of
Industrial
Experiments,
Hafmer
Publishing
Company,
New
York.
7.
Jones,
R.
P.,
N.
Pammet
and
P.
F.
Greenfield,
1981.
Alcohol
fermentation
by
yeasts
the
effect
of
environmental
and
other
variables,
Process
Biochemistry
16
(3):
42-9.
8.
Martinez,
G.
and
N.
Murphy,
1984.
Congeners
from
high-test
molasses
alcoholic
fer-
mentation,
J.
Agric.
Univ.
P.
R.
68
(1):
59-66.
9.
Murphy,
N.
F.,
1983.
Annual
Progress
Report
of
Research
Project
H-279,
Factors
affecting
the
fermentation
efficiency
of
molasses
by
the
yeast
and
their
effect
on
the
quality
of
rum,
Rum
Pilot
Plant,
Agric.
Exp.
Stn.
Univ.
P.
R.,
Rio
Piedras,
P.
R.
10.
Murphy,
N.,
1984.
Fermentation
of
high-test
molasses,
J.
Agric.
Univ.
P.
R.
68
(1):
33-44.
11.
Nykanen,
L.
and
H.
Soumalainen,
1983.
Aroma
of
Beer,
Wine
and
Distilled
Alcoholic
Beverages.
D.
Reidel
Publishing
Company,
Dordrecht,
Holland,
p.
4.
12.
Official
Analytical
Methods
of
the
Rum
Pilot
Plant,
1969.
Agile.
Exp.
Stn.
Univ.
P.
R.,
Rio
Piedras,
P.
R.
13.
Ramirez,
M.,
1982.
Characterization
of
slops
of
high
test
molasses
alcoholic
fermenta-
tion,
J.
Agric.
Univ.
P.
R.
66
(3):
235-37.
14.
Reich,
G.
T.,
1945.
Molasses
distiller's
dried
yeast,
Sugar
40
(1):
26-9.
15.
Rosado,
E.,
1982.
Metodo
para
la
elaboration
de
melaza
rica,
Progress
Report
No.
23-83,
Rum
Pilot
Plant,
Agric.
Exp.
Stn.
Univ.
P.
R.,
Rio
Piedras,
P.
R.
16.
Rose,
A.
H.
and
J.
S.
Harrison,
Ed,
1971.
The
Yeasts,
Volume
2,
Academic
Press,
London
and
New
York,
pp.
4-5.
17.
Siegel,
S.,
1956.
Nonparametric
Statistics
for
the
Behavioral
Sciences,
McGraw-Hill
Book
Company,
New
York.
18.
Torres,
C.
S.,
Z.
de
Ayala
and
M.
I.
Rivera,
1987.
Aging
of
high-test
molasses
rums.
J.
Agric.
Univ.
P.
R.
71
(2):
137-58.