1. Introduction: - ASSBT Proceedings

Chemistry of Storage Problems Encountered with Molasses Desugarization Extract
By David R. Groom, Terry D. McGillivray, James H. Heggeness, and Indrani S. Samaraweera
American Crystal Sugar Company, Technical Services Center,
PO Box 1227, Moorhead, MN 56561-1227
1. Introduction:
American Crystal Sugar Company (ACSC) has two molasses desugarization(MDS) facilities. It is the
practice of ACSC to store the extract juice from each desugarization facility for up to nine months, then
process the stored extract juice after completion of the beet processing campaign. Extract juice has been
stored and processed with little incident at our East Grand Forks (EGF) MDS facility commissioned in
the fall of 1993. The EGF facility is a "standard" simulated moving bed (SMB) separator.
In January of 2000 ACSC commissioned a second MDS facility at the Hillsboro (HLB) factory location.
The separator utilizes displacement chromatography along with simulated moving bed technology (1).
After resolution of initial start up problems, (cell packing, cell repairs, control system issues, etc.) the
separator was producing 92-93 purity extract and the extract was sent to storage. Shortly thereafter,
extract purity dropped. Through the course of several months average purity of the extract dropped to
90; with purities as low as 88.5 observed.
When the extract was brought into the sugar end for processing, numerous problems were encountered
that lead to poor recovery. Recovery was 5.3 cwt per ton standard molasses versus the expected 7.64 cwt
per ton. Some of the poor recovery was attributed to the poor quality extract produced during
commissioning of the MDS system. Raffinose concentration in the feed was up to 3.0% on solids and up
to 1.4% DS in extract. Invert sugar levels were also elevated. Invert concentrations as high as 0.3 % DS
were observed in the initial extract produced; whereas invert levels in stored extract processed were as
high as 2.0 % on DS.
Table I
2000 Extract Campaign Results
Extract sent to storage Extract as processed
88.5 to 91.5 "true" purity 88.0 to 91.0 "true"
10,000-12,000 color 10,000 to 30,000 color
9.0 to 9.3 pH 7.5 to 9.0 pH
7.64 cwt/ton molasses sucrose recovery target 5.3 cwt/ton ;sucrose recovery
Invert 0.3 % on DS Invert up to 2.0% on DS
Additional work was completed on the separator with the goal of improving performance. The effort
achieved lower color, higher purity, lower invert, and decreased raffinose level in the extract sent to
storage. Extract purity and recovery improved from September 2000 through May 2001.
Pilot plant work done on the extract completed in months prior to the second extract campaign produced
good recovery of sugar. However, during the second extract processing campaign, many of the same
problems plagued the sugar end that had been experienced in the prior campaign. Recovery was low at
6.1 cwt/ton.
Further investigation into the storage and production of extract was initiated. Many of the quality
measurements that had been used to monitor EGF extract in storage did not provide adequate assessment
of the Hillsboro extract. Extract quality indicators such as lactic acid, pH, apparent purity, RDS,
microbial counts standards used for EGF product, seemed inadequate. Storage parameters for Hillsboro
\.
3

4
extract had been set initially to those of EGF product; RDS of 68.S, temperature below 2SoC, pH target of
nme.
Table 2
2001 Extract Campaign Results
Extract sent to storage Extract as processed
90.0 to 94.0 IC purity* 89.S to 93.S IC purity* \.
8000-9000 color 8000 to 20,000 color
9.4 to 9.6 pH 9.1 pH
7.64 cwt/ton molasses sucrose recovery target 6.1 cwt/ton sucrose recovery
Invert 0.1 % on DS Invert up to 2.9% on DS (1 tank)
*IC purity refers to purity determined by ion chromatography.
In reviewing data, it was found that invert levels in the extract had increased in storage. Invert levels
were found to be higher in the Hillsboro extract entering the factory than when the extract went into
storage. Since high color was a major issue in processing the extract, we expressed the color on
nonsugars as described in Sugar Technology (2). The Hillsboro extract color to nonsugar (CINS) ratio
was found to be higher than normally encountered in previous extract processing at EGF or in HLB thick
juice. Additional tests indicated that water activity and buffering capacity of the Hillsboro extract were
different than the EGF extract. Biocides such as formaldehyde, hops beta acids, and sulfur dioxide had no
impact on stored extract stability. Investigation into the extract processing/storage issue proceeded along
several paths.
In order to evaluate conditions required to store Hillsboro extract, a comprehensive study was conducted
to look at the impact of various variables on the stability of the Hillsboro extract. An accelerated storage
study was completed using coupled loop extract from Hillsboro and extract from our East Grand Forks
facility. Samples were incubated at 50°C to accelerate reaction rates. Hillsboro extract was observed to
decrease in purity and pH with an increase in color and invert with little change in microbial counts or
lactic acid. Similar results had been observed in the tanks but occurred over a longer time frame.
Ion chromatographic data indicated that concentration of invert was correlated with the degree ofjuice
degradation. Increased concentration of invert in the stored extract was not correlated with increased
microbial counts, lactic acid or other signs of microbial infection (3). Based on the prior mentioned
observations, we investigated other causes for juice degradation including chemical and enzymatic
activity. Invert sugar as determined by IC was found to be an effective predictor ofjuice degradation (3).
Invertase activity was found in the pH range of 8 to 10.0, negligible at pH 1O.S, no activity noted at pH
11. High pH and pasteurization were found to mitigate degradation.
Recommendations to monitor invert in production and storage extract gave us sufficient early warning to
process a tank in time to avoid significant loss. Even though microbial counts in the stored extract didn't
seem high enough to indicate problems, tracking invert allowed us to indirectly monitor enzymatic
(invertase) activity. High nitrite levels that appeared in the extract evaporator loop indicated problems
with the evaporators. In our studies, elevated nitrite was associated with unstable extract. Elevated micro
counts were found in the extract evaporator loop. Interestingly, microbial counts did not always correlate
to high nitrite levels (4). It was suggested that the temps (SO-SS°C) in the extract evaporators were ideal
for certain microbes to proliferate. A recommendation was made to raise the temperature in the extract
evaporator loop to a minimum of 7S°C. In addition, the pH of the extract prior to evaporators was raised
to 10.5. Microbial and nitrite data indicated a decrease in microbial activity in the extract. Increased
temperature in the water loop decreased sliming of check filters, indicating a reduction in microbial load
to the trains.

2. Materials and Methods:
Sample collection: Samples were collected from various points in the MDS processing facility, extract
storage tanks, and sugar end stations. Purging of sampling ports/valves was done before collection of
samples. Samples were collected using appropriate aseptic method and placed in sterile containers. All
subsequent handling of samples was accomplished using sterile techniques. Accelerated storage test
samples were placed in an incubator and kept at 50°C for the duration ofthe test period. Sub samples
were taken periodically from the accelerated storage test containers via sterile pipette for analytical and
microbial analysis.
Analytical methods, chemical analysis: Wet chemical analysis was completed'using standard ICUMSA
methods. Determination of pH was done by using a Ross combination electrode and an Orion pH meter.
Apparent purity was determined on a Rudolph AutoPol 880 using 880 nm wavelength and a 10 cm cell.
RDS was determined on a Bausch and Lomb refractometer. Color was determined by measuring
absorbance on a Milton Roy Spectronic 21 D using a 1 cm cell. Samples were adjusted to pH 7 before
absorbance was read. Water activity was measured with an Aqualab CX2; 0.78 N NaCl was used as a
calibration standard. Invert (glucose and fructose) and sucrose concentrations were determined on a
Oionex OX 500 Ion Chromatograph (IC) unit, using a Carbopac PAl colulllil an ED 40 detector in the
PAD mode. The system was operated in isocratic mode at 22°C. Eluent was 200 millimolar sodium
hydroxide kept under pressurized helium. Acetic, butyric, formic, lactic, and propionic acids were
determined on Waters 500 series HPLCs using refractive index detection in isocratic mode. Eluent for
the HPLC was a 5 millimolar sulfuric acid mobile phase; a BioRad Aminex HPX-87H colulllil was used
at 35°C. Betaine was determined with the Waters HPLC using a Bio-Rad Aminex HPX-87N colulllil,
eluent at 0.6 mllminute 10 millimolar sodium sulfate and refract index detector. Nitrite concentration was
estimated by using EM Quant test strips, samples were diluted with de-ionized water before reading.
Microbial analysis included: Mesophilic, thermophilic, regular and osmophilic yeast and mold counts.
In addition, flat sours, total thermophilic spore, along with thermophilic anaerobes producing H2S, and
thermophilic anaerobes not producing H2S were counted. The methods used are described in detail by
Samaraweera et al (4).
Accelerated Laboratory Storage Testing:
Accelerated storage studies were started in July 2001 and continued in various form until December 2002.
Most of the studies ran from 6 to 10 weeks. All accelerated storage trials (AST) involved collecting
extract from tanks or production extract to tanks and placing in sterile HDPP 1 'or 3 liter containers.
Samples were then incubated at 50°C. Aliquots were drawn for chemical analysis at regular intervals.
Aliquots were also drawn from AST samples on selected trials and microbial analysis completed. Details
of the microbial results are addressed by Samaraweera et al. (4). Method details of various trials
conducted will be addressed in Results and Discussion section of this paper.
3. Results:
Trial 1: Involved collection of production extracts on July 19, 2001 from EGF and HLB. Samples were
placed in 3 liter containers and stored at 50°C. In addition, two 3 liter samples were stored at ambient
temperature. The test ran for 10 weeks. Samples incubated at 50°C will be referred to as the AST
sample.
At the end of the 10 week trial period, there were obvious differences noted between the EGF and HLB
extract AST samples. Purity as measure by IC dropped 0.6 points in the EGF extract while the HLB
extract purity dropped 4 points as shown in figure 1. Both ambient samples had little change in purity.
As shown in figure 2, the HLB extract AST sample increased in color from 7700 to 24,000. The EGF
AST sample color increased from 7600 to 14,000. Note in figure 3 that the pH dropped significantly in
the HLB AST sample, from 9.6 to 7.l. Figure 4 shows that invert in the HLB AST sample increased 16
5

­
% Retain AP
X = A: Purity HLB
Y = B: pH HLB
Actual Factor
C: RDS= 70
% Color rise
X = A: Purity HLB
Y = B: pH HLB
Actual Factor
C: RDS= 70
188
173
157
% Color n58 142
126
B: pH HLB
11.0
10.5
10.0
Figure 6
9.5 95.5
9.0 95 .0
Figure 7
B: pH HLB 9.5 95 .5
9.0 95 .0
96.5
A: PurityHLB
96.0
96 .5
A: Purily HLB
97.0
9

. . _ ._­
14
Note the predicted response for color below. Color is the highest at the highest invert level, which
occurred at the lowest pH (9). At pH 11, invert increase was low, regardless of the invertase addition.
Invert increase in the samples were also minimal when no invertase was added regardless of the pH. For
this trial, high pH treatment clearly stopped invertase activity.
Color
X = A: pH
Y = B: Invertase 30772
25616
20464
Color 15310
10157
10.0
% Invert
X = A: pH
Y = B: Invertase 20.9
% Invert
15.7
10.4
B: Invertase
5.2
0.0
100
0.0 9,0
Figure 15
Figure 16
B: Invertase 2.5 9.5
0,0 9,0
10.5
A: Alkalinity
A: pH

Trial 10: In Trial 10, the focus was on the impact/interactions of invert and pH on the stability of
Hillsboro extract (6). The HLB extract was autoclaved, and then treated with acid or base to adjust pH.
In addition, a sterile preparation of invert sugar solution was added to the extract. Levels of invert were
0.02 to 0.07% on extract after autoclaving. Invert was adjusted to targets of 1 and 2% on sample as
required at the start of the trial. Extract pH was also adjusted; the initial targets were 9, 10, and 11.
A slight IC purity drop of 0.2 to 0.5 units was noted for the treatments. Interesting differences were found
for other measured responses as noted in Table 7 below. At low invert level, color was unaffected by pH.
As pH and invert level increased, color increased. Color increase was the bighest at high pH and high
invert level. The greatest increase in acetic acid was observed under conditions of high pH and invert. At
low levels of invert, acetic acid increased little regardless ofpH. At low pH, acetic acid level was less
influenced by invert level. Lactic acid concentrations varied little between treatments.
Table 7
Treatments Initial Data Ending Data
Trial # 10 Color % Invert Acetic ppm Color % Invert Acetic2pm
0% Invert, pH 9 10,370 0.020 342 13,140 0.008 388
2% Invert, pH 9 10,060 1.987 331 28,160 1.158 988
0% Invert, pH 11 10,000 0.012 350 11,470 0.008 392
2% Invert, pH 11 9580 1.773 331 46,310 0.018 2362
Discussion:
The major issues regarding extract were increased color in storage, high color in sugar end processing,
low buffering capacity, decreased purity in storage, and increased invert levels.
Early in the campaign 2001 recommendations were made to keep RDS above 69.5, the pH at 9.5, and
temp to storage less than 20°C. Sargent indicated that stability was noticeably improved ifjuice was
stored at 10°C (3, 7). With improved cooling equipment we were able to target and hold 15°C to storage.
To prevent microbial loading into the system finer filter media was used for filtering the feed molasses.
Recommendations were considered the usual target values for juice storage in the sugar industry. While
these measures may have helped stabilize the extract somewhat, they were insufficient to maintain desired
extract quality in storage. Further investigative work was required to address the storage issue.
Through all the testing, East Grand Forks extract was more stable (chemically) than Hillsboro extract.
Various treatments to improve stability of Hillsboro extract were not effective. The ineffective treatments
included beta acid hops, sulfur dioxide, and formaldehyde.
Buffering capacity of the Hillsboro extract was found to be 2 to 5 times lower than EGF extract. As the
degradation reaction occurred, a lack of buffering capacity led to an accelerated pH drop which in turn
increased enzymatic, microbial, and chemical degradation. In our studies, juice that was buffered did not
degrade as rapidly. Based on test results, a recommendation was made to raise the pH of the extract sent
to storage. Raising the pH to 10.5 also retarded enzymatic and microbial activity. In addition, the action
increased buffering capacity of the extract. While the course of action did not address the root cause, the
juice was stabilized. Invert levels were held in check with the high pH treatment. Also to lower the rate
of reactions a recommendation was made to lower the target temp ofjuice to storage to 15°C or less.
Since water activity impacts rates ofchemical and biological processes, we tested stability of Hillsboro
extract at a range of RDS from 68 to 72. In our studies we did not see any significant difference between
extract stored at 70 or 72 RDS. There may have been some contribution to the problem when extract was
sent out to storage at 68 RDS. The recommendation to elevate the RDS to above 69.5 reduced water
activity impact.
15

-----'"
18
7. Sargent, D.; Briggs, s.; Spencer, S. (1997): Thickjuice degradation during storage. Zuckerindustrie. 122, Nr.
8, pp. 615-621
8. Bruijn, de, 1.M: Kieboom, A.P.P.; Bekkum, van, A. (1986): Sugar Teclmology Review 13, pp 21-52.
9. Bruijn, de, 1.M; Struijs, 1.; Bout-Diederen, M (1998): Sugar degradation and colour fonnation. Proceedings,
SPRl Conference. Savannah, GA, USA pp 127-143.
10. Madsen, R.F.; Nielsen, WK. ; Winstrom-Olsen, B.; Nielsen, TE. (1978-79): Fonnation of colour compounds in
production of sugar from sugar beet. Sugar Teclmology Reviews, 6, pp 49-115.
II. Godshall, M A. (1992): Isolation of a high molecular weight colorant from white beet sugar. Proceedings, 1992
Sugar Proc. Res. Conf., pp 312-323.