Glycerol production in industrial alcohol fermentations

Biotechnologg Letters Vol I0 No 9 Received August 18

677-682

(1988)

GLYCEROL PRODUCTION IN INDUSTRIAL ALCOHOL FERMENTATIONS P.J. Brumm* and R.E. Hebeda Enzyme Bio-Systems Ltd. 3350 Salt Creek Lane Arlington Heights, IL 60005 Glycerol production in fuel ethanol fermentations was shown to be related to yeast dosage and initial dextrose concentration. Use of a partially saccharified (30 D.E.) syrup in place of dextrose resulted in lower glycerol production; the 30 D.E. syrup also reduced the effect of yeast dosage on glycerol production. The results obtained suggest that a portion of the glycerol produced during alcohol fermentaions is for NADH regeneration rather than for osmotic stabilization. INTRODUCTION For the fuel ethanol fermentation, starch (usually corn starch in the U.S.) is first liquefied with an alpha-amylase and then saccharified with a glucoamylase to produce a highly concentrated (usually ~30% by weight ) dextrose-rich feedstock. This feedstock is then fermented in a complex medium, and the alcohol is recovered by fermentation (Keim, 1983). Glycerol is an unwanted by-product of the fermentation. Glycerol production by the yeast uses up valuable substrate that normally would be converted to ethanol. The glycerol produced is at a very low concentration; recovery of it is not economical. Since the glycerol cannot be recovered, it ends up in the stillage stream; if the stillage is dried for feed production, the hydroscopic nature of glycerol increases drying costs substantially. Glycerol production by yeast has been studied (Vijaikishore and Karanth, 1986) often with the goal of increasing production to levels high enough to justify cornmercialization. Glycerol production by yeast has been linked to osmotic stabilization of the organism. The relevance of this work, which utilized high levels of salt (Edgley and Brown, 1983, Maiorella, et al, 1984) or non-fermentable carbohydrates (Ziffer, 1983) to increase solution osmolarity may not be relevant to conditions experienced in industrial fermentations. This work examines the potential causes of glycerol production in fuel ethanol fermentations and possible solutions to the problem. MATERIALS AND METHODS

Corn starch (Code 3005) and anhydrous dextrose (Code 2401) are products of the Corn Products Division of CPC International, Englewood Cliffs, NJ. Enzymes, G-995 Alpha-Amylase and G-990 Glucoamylase, are products of Enzyme Bio-Systems, Ltd, Englewood Cliffs, N.J. and are dosed based on labeled activity. Frozen samples of light stillage and light corn steep liquor were obtained from Pekin Energy Company, Pekin, IL. Dry Active Distillers Yeast, Saccharomycescerevisiae, (DADY) was obtained from Universal Food Products, Milwaukee, WI. Methods A corn starch slurry (30% on a w/w basis) was adjusted to 0.8 mM CaCI2 (100 ppm calcium on a dry substance basis), and the pH of the slurry was adjusted to 6.2 by addition of 1M NaCO3. The slurry was then dosed with 4.0 Enzyme Bio-Systems alpha-amylase units/gram of starch, and the starch was liquefied by incubation at 98~ C for 90 minutes. The liquefied starch was cooled to

677

60 ~ C, and adjusted back to original weight to maintain a constant dry substance. The pH was adjusted to 4.2 to 4.3 by addition of 1 N HCI, and 0.006 Enzyme Bio-Systems glucoamylase units per gram of starch was added. The liquefact was then incubated for 24 hours in a 60 ~ C bath; the partially saccharified material was adjusted back to its original weight and then immediately used for fermentation medium preparation. The dextrose equivalent of this material was 28-30 D.E. as determined by HPLC (Bernetti, 1982). The fermentation medium contained 10% (w/v) light. (unevaporated) corn steep liquor, 30% (w/v) light (unevaporated) stillage, and 8% to 20% (w/v on a dry basis) carbohydrate. Carbohydrate concentrations of the 30 D.E. syrups were corrected for chemical gain (the weight gain in carbohydrate due to the addition of water) by multiplying the initial starch weight by 1.11 (Hebeda, 1987). The medium was adjusted to pH 5.2 with NH4OH and diluted to final volume by addition of water. The medium was used without sterilization; immediately after adjustment of pH and volume, dry active distillers yeast (0.5 g/100 ml of medium, except in yeast dosage experiments) was added to initiate the fermentation. To fermentations containing 30 D.E. syrup, supplemental sterile glucoamylase (filtered through a 0.22 I~ filter) was added at a dosage of either 0.15 or 0.25 units/gram of carbohydrate. Fermentations were run at 28 ~ C in 125 ml flasks containing 100 ml of medium with 100 rpm agitation on a linear shaker. All fermentations were performed in duplicate and all individual samples were assayed in duplicate. Samples were filtered through Whatman #1 filter paper to remove yeast cells and solid material. The filtrates were assayed for solvent production and residual carbohydrates by HPLC. Separations were performed on a 1-foot HPX-87 column in the hydrogen form (BioRad Laboratories, Richmond, CA) using 0.06 N H2SO 4 as eluent. Eluted compounds were detected using a differential refractometer and quantitated using an electronic integrator and gravimetrically prepared standards for calibration. HPLC indicated that only ethanol and glycerol were produced as fermentation products under these conditions. RESULTS The ratio of glycerol/ethanol in completed fermentations was measured as a function of the initial yeast dosage. When 20% dextrose was used as substrate, glycerol production ranged from 0.10 to 0.17 grams of glycerol per gram of ethanol. Glycerol production was minimized at a yeast dosageof approximately 0.5 g dry yeast per 100 ml of medium (Figure 1). Increasing or decreasing the yeast dosage 10-fold resulted in a greater than 50% increase in the amount of glycerol produced. Use of a 20% concentration of the 30 D.E. syrup (supplemented with 0.25 glucoamylase units per gram of carbohydrate) resulted in a significant decrease in glycerol production. At all yeast dosages tested, the 30 D.E. syrup fermentations gave lower glycerol/ethanol ratios than did the dextrose fermentation at its minimum. The fermentations using the syrup had a minimum glycerol production at a lower yeast dosage (0.2 g per 100 ml) than did the dextrose fermentations; at this yeast dosage, the glycero! produced was at least 20% lower than the minimum glycerol production seen with dextrose as substrate. Using the 30 D.E. substrate, much smaller variations were seen in the amount of glycerol produced when the yeast dosage was varied; raising or lowering the yeast dosage 10-fold resulted in only a 20% increase in glycerol formation. The ratio of glycerol/ethanol in completed fermentations was measured as a function of the initial dextrose concentration, using a constant yeast dosage of 0.5 g/100 ml. Based on osmotic considerations, one would expect an increase in glycerol production with increasing substrate concentrations. Increasing the dextrose concentration from 8% to 20% resulted in an increase in glycerol production from 6.0 g/L to 11.3 g/L. This increase in glycerol production is less than the relative increase in ethanol production: as a result, the important glycerol/ethanol ratio actually decreases 27% when the dextrose concentration is increased (Figure 2) over the range of 8% to 20%.

678

FIGURE 1 EFFECT OF YEAST DOSAGE ON GLYCEROL PRODUCTION 0.18

A

0.16

-J 0 z ,<

0.14

p. LU _1 O m I.U O >. ,.I (.-1

0.12

o--

30 I~.E. S Y R U P 0.10

0.08

0.06

........

, .1

.01

........

, 1

Y E A S T D O S A G E [g/100 ml]

FIGURE 2 EFFECT OF DEXTROSE CONCENTRATION ON GLYCEROL PRODUCTION 0.18 A

0.17 .J 0 z ,< Iw .-I 0 nLu 0 >. .J

DEXTR(~E

\

0+16

0.15 0.14

0.13

0.12

I

6

10

i'8

1'4 % DEXTROSE

679

22

0

___

~-

Z

Z

?

,_1 0

0

Z

< 3: u) I-LUU.I

u.i

u.i,

d o

".."

.

d o

1i

Ill

I~.o

o -

~ 0 ILUl r

,;

i

~

~

.J

! in

~

W

~.o-"

~*~

:

:

p ' ". ~ /

L

...~:.:~"

P

0

. . . . . ....~:..:....cr" ......

I 0

I If) 1,,-

0

i C3

0

d

0

[o/o]

HOJ.~/'IOU~OA'IO

<

< 0 LO

Z

_o

it)

I0

LN

0

0 CO

0

rr

lil~. I ~ ,..i ~0 _~,,. u.-I (3

.~ "l=

9

~.~ ii " ' a

,

i

0

I,I i>! ~!

o UJ

0 ,r--

I--

.J !

!

I

i

[1/0]

Z

_o

0

!

IOU:;OAIO <

<

9a

I-~

i--

i\

u~

m

O U')

O ,q-

F-

~ e0 0 met flCO.

< -r t-u.I

d

=

!

!

i

i