Patent Application
20050214408
1. Field of the Invention
The present invention relates to the production of potable alcohol products, especially beer, and in particular using a hybrid process comprised of continuous and batch fermentation processing stages.
2. Description of Related Art
The extensive number of recent publications in this area illustrates the brewing industry's great interest in immobilization. In several reviews (Enari, 1995; Iserentant, 1995; Masschelein, 1997; Mensour et al., 1997; Stewart, 1996; Virkajarvi & Linko, 1999) on the general state of the brewing industry, the possible revolutionary role of immobilization in beer production has been highlighted. In addition to the groups mentioned in the following sections, many other institutions have been involved in immobilized cell R&D for brewing applications. The Miller Brewing Company (Duncombe et al., 1996; Tata et al., 1999) from the United States ran some preliminary evaluations on a Meura Delta test unit, as well as on the fluidized bed bioreactor distributed by Schott Engineering. Coors Brewing also performed preliminary experimentation with the Meura Delta system.
The Slovak Technical University, in collaboration with Heineken, investigated the use of calcium pectate gel in an air lift system for the production of beer (Domeny, 1996). More recently, the group from the Slovak Technical University has published several articles on their on-going research (Smogrovicova et al., 1997; Smogrovicova & Domeny, 1999). Guinness initially investigated the use of various adsorption carriers for immobilization and subsequent fermentation in a fluidized bed bioreactor (Donnelly, 1998). In 1999, Donnelly and colleagues published a paper describing the kinetic of sugar metabolism inside their fluidized bed bioreactor (Donnely et al., 1999). Their experimental setup involved the use of SIRAN porous glass beads as the immobilization carrier for a top fermenting yeast.
Holsten Brauerei AG and Lurgi AG, both from Germany, have jointly developed and operated a pilot plant for the continuous production of alcohol free beer (Dziondziak, 1995). Calcium alginate beads within a single-stage loop fluidized bed fermentor (130 L) were used for fermentation while a sieve bottom column (seven sieve bottoms) was used to dealcoholize the beer. The total production time for this process was 8.5 hours.
A team from the Sapporo Breweries Ltd. Brewing Research Laboratories located in Japan studied the use of immobilized cells in a fluidized bed reactor for the main fermentation of beer. Their studies involved the use of polyvinyl alcohol gel beads (Shindo & Kamimur, 1990), Ca-alginate gel beads (Shindo et al., 1994a), double-layered gel fibers (Shindo et al., 1994b) and chitosan gel beads (Shindo et al., 1994c) as immobilization matrices. In the latter study, a one liter working volume bioreactor containing 25 percent by volume CHITOPEARL® type II beads (chitosan beads) was operated on a continuous basis with wort treated with glucoamylase. This enzyme treatment allowed for acetate ester formation in the immobilized cell system to be similar to that of conventional batch fermentation, thus one step closer to product matching.
The research group from Sapporo has now focused its attention on the development of a chitosan bead fluidized bead fermenter operating in repeated batch mode. The system was operable for 75 days without any major problems and the resulting beer was similar in quality to a commercial product. A non-flocculating strain was shown to be much more effective than a flocculent strain (Maeba et al., 2000; Umemoto et al., 1998).
Two other approaches to diacetyl reduction were presented at the EBC Congress held in Maastricht in 1997. Researchers in France (Dulieu et al., 1997) proposed the use of encapsulated α-acetolactate decarboxylase to rapidly convert α-acetolactate into acetoin. Meura Delta has highlighted preliminary results on the use of an aluminosilicate zeolite as a catalyst for the cold and direct conversion of acetolactate into acetoin (Andries et al., 1997). If such treatments prove to be effective and consumer acceptable, lower cost alternatives to the maturation systems proposed by Cultor and Alfa Laval could become a reality.
Other non-industrial research in the field of immobilization for beer production includes that of the Singapore Institute of Standards and Industrial Research where the use of a thread type of alginate gel particle for use in a packed bed reactor was studied and found to be more favorable than alginate beads (Que, 1993). Mafra and colleagues from the Universidade do Minho in Portugal have discussed the use of a superflocculent yeast strain for the continuous maturation of beer (Mafra et al., 1997). This same research group has also published work on the use of their flocculent yeast within an airlift bioreactor for the production of ethanol (Vicente et al., 1999; Domingues et al., 2000).
Researchers at various academic institutions have also recently investigated the use of immobilization for the production of beer (Argiriou et al., 1996; Bardi et al., 1996; Cashin, 1996; Moll & Duteurtre, 1996; Nedovic et al., 1996a; Nedovic et al., 1996b; Norton et al., 1995; Scott et al., 1995; Wackerbauer et al., 1996a; Wackerbaue et al., 1996b). Research groups from China (Chao et al., 1990; Yuan, 1987; Zhang et al., 1988), Russia (Kolpachki et al., 1980; Sinitsyn et al., 1986) and Czechoslovakia (Chladek et al., 1989; Curin et al., 1987; Polednikova et al., 1981) were also involved in immobilized cell technology and published results in the 1980s.
The present invention relates to a process for the production of potable alcohols, which comprises a continuous fermentation stage that is employed to pitch and/or at least initially ferment a wort containing fermentable sugars.
In particular a preferred process is provided in which the continuous fermentation is carried out using a gas lift type bioreactor, employing a flocculent (and especially a highly flocculent or superflocculant) yeast strain and employing stringent oxygen control. In a particularly preferred form of the present invention, the “at least partially fermented” discharge from the continuous process is delivered to a batch processing stage for finishing (which in the context of the claims of the present invention can include—but is not limited to—the completion of the fermentation process through which fermentable carbohydrates to alcohol).
The present invention relates to a process for the production of beer (including in particular pale styles of beer, lagers, and especially North American style beers) wherein the preferred form is a continuous stage carried out in a gas lift bioreactor.
In accordance with a continuous stage useful in the various practices under the present invention, it is preferred that immobilized cells be utilized, (as opposed to purely free cells) and this may be carried out using a selected one of carrier immobilized or flocculating yeasts. Notwithstanding the forgoing, it is preferred that flocculating yeasts be used instead of carrier immobilized cells, and superflocculant yeasts are especially preferred for this purpose.
More details concerning preferred practices and advantages associated with the continuous processes are provided over the course of the detailed description of the present invention. These include the use of artificial (e.g., controlled) gas mixtures and the use of nitrogen, carbon dioxide and oxygen as well as air.
In addition, greater details concerning the batch hold processing stage are provided herein. Note that in certain embodiments of the present invention, the focus of the batch hold process goes beyond issues of “completion” of the conversion of fermentable carbohydrates to alcohol (which in any case can be virtually completed in the continuous stage of the processing). In such embodiments, the primary focus of the batch hold processing stage is on flavor-matching (or remediation), particularly in connection with diacetyl and acetaldehyde.
Preferred embodiments of the present invention provide for the post-continuous stage distribution of the pitched and/or at least partially fermented wort through a distribution manifold (whether as a fixed manifold or by selectively connecting and disconnecting conduit) amongst a plurality of batch hold tanks. In a serial distribution process one tank is filled, followed by the next, and so on. In a particularly preferred embodiment, the continuous reactor throughput capacity and batch hold capacity are matched in terms of size and number of reactors/batch hold vessels—such that the production flow-rate is matched in terms of capacity over time. Ideally, a batch hold vessel is drained of finished product just in time to be cleaned, reconnected and then refilled from the ongoing discharge from the continuous fermentation stage.
In accordance with another aspect of the present invention, certain embodiments are particularly managed in relation to the oxygen content of the wort/beer. This applies to both the continuous and batch hold stages of the process. With regard to the continuous stage the oxygen concentration has a variety of effects, but notably, it may be desirable to minimize it in order to optimize conversion of higher alcohols to flavor active esters. In this connection it is noted that concentrations of higher alcohols can remain largely unaffected by the batch hold processing stage, so if desired stringent O2 control is used to manage the fusel ester flavor balance. Pre-purging of wort with CO2 prior to continuous fermentation can be useful in this connection.
In one embodiment of the present invention, the primary purpose of the continuous stage of the processing is to provide for pitching of the downstream batch fermentation that then occurs in the batch hold process.
For greater certainty, the contents of the priority documents are incorporated herein in full and form as much a part of the present specification as if they were fully reproduced herein.
The following is a two-part detailed description of aspects of the present invention.
Detailed Description—Part 1:
Yeast Strain and Inoculum Preparation
The fermentations conducted in this invention employed a polyploid yeast from the Saccharomyces cerevisiae family (also referred to as Saccharomyces uvarum and/or Saccharomyces carlsbergensis). The brewing community will commonly refer to this yeast as bottom fermenting producing a lager-type beer. This characterization is attributable to lager yeast's ability to settle out of the liquid medium upon completion of the fermentation. Ale yeast, unlike the lager yeast, will rise to the top of the fermentation vessel and was therefore known as a top fermenting strain. The ability of yeast to settle or rise is not necessarily dependent on whether the yeast is a lager or ale type but is strain specific. Lager yeast typically does not ferment at temperatures above 34° C. while ale yeast cannot ferment melibiose. Scientists will use these characteristics to differentiate lager strains from ale yeast (McCabe, 1999).
The medium flocculent yeast strain Saccharomyces cerevisiae strain 3021 from the Labatt Culture Collection was used in both the free cell self-aggregated fermentations and the κ-carrageenan immobilized fermentations. For the trials involving the use of superflocculent yeast as the immobilizant, a variant of the LCC3021 strain, namely LCC290, was used.
Pure yeast cultures were cryogenically stored in a −80° C. freezer located within the Labatt Technology Development Department. When required, sterile loops of yeast culture were aerobically pre-grown at 21° C. on PYG agar plates (3.5 g of peptone, 3.0 g of yeast extract, 2.0 g of KH2PO4, 1.0 g of MgSO4.7H2O, 1.0 g of (NH4)SO4, 20.0 g of glucose, and, 20.0 g of agar dissolved in distilled water up to a volume of one liter). Isolated yeast colonies were then transferred into test tubes containing 10 mL of pasteurized wort and incubated with agitation at 21° C. for a 24 hour period. This inoculum was progressively scaled up to a volume of 5 L by adding the previous culture into the appropriate wort volume (10 mL into 190 mL, 200 mL into 800 mL and 1 L into 4 L). The yeast inoculum was then transferred into centrifuging jars and subjected to centrifugation at 10000 rpm and 4° C. for 10 minutes. The desired mass of yeast for all subsequent fermentations was drawn from the resulting wet yeast pellets (30 percent w/v).
Fermentation Medium
Industrial grade lager wort produced by the Labatt London brewhouse was used as the nutrient medium for all fermentations. Reference is made throughout this invention to the wort's specific gravity expressed as degrees Plato (° P). The following formula describes the relationship between specific gravity and ° P.
° P=135.997·SG3−630.272·SG2+1111.14SG−616.868
The wort used was 17.5 ° P which corresponds to a specific gravity of 1.072.
Table 1 provides the typical carbohydrate profile of this wort as measured by the high performance liquid chromatography (HPLC) method described in section 4.7.2. Approximately 73 percent of the carbohydrates in this wort are fermentable, while the brewing yeast employed in this study cannot readily take up 27 percent of the longer chain carbohydrates.
The coefficient of variation for most of the analyzed substances ranges between 10 percent and 20 percent. This variability is due in large part to the industrial production process utilized, as well as the variability in raw materials from one brew to another.
Immobilization Types
Three types of immobilization—entrapment, adsorption, and self-aggregation—were tested during this invention. For the industrially sourced carriers, supplier data are presented first and then supplemented by in-lab analysis. Pictures and size distributions of the investigated carriers (when available) are presented elsewhere herein.
Two types of adsorption matrices were tested in the pilot scale gas-lift draft tube bioreactor. Pictures of both these carriers are presented herein. Schott Engineering provided a sintered glass bead carrier, SIRAN®. The selected particles were one to two mm in diameter, had open pores for yeast immobilization with a 55 to 60 percent pore volume, and pore size distribution between 60 and 300 μm, an appropriate size for yeast cells. This type of carrier is reported to be biologically and chemically stable, easy to clean, reusable, sterilizable with steam, non-compacting and neutral in taste, and is therefore food approved.
World Minerals of California supplied a spherical carrier composed of diatomaceous earth. This carrier provided the advantages of thermal and chemical stability, mechanical strength and rigidity. Diatomite, the basic raw material, is commonly used in the brewing industry for filtration of beer. The CELITE® R-632 carrier was specifically designed for whole cell immobilization.
Supplier specifications were as follows:
Kappa-carrageenan gel beads, an entrapment based carrier, were produced in the laboratories of the Labatt Brewing Company Limited. The production process and the results are presented below.
The simplest mode of immobilization, self-aggregation, was possible by the selection of yeast strains capable of flocculation. The industrial lager yeast LCC3021 possesses the natural capability of flocculation and is considered as a medium flocculent strain. As the fermentation progresses, small clumps of yeast, measuring from 0.5 mm to 1.0 mm, will form in the liquid medium. The LCC290 yeast, a variant of LCC3021 lager yeast, will form much larger flocs (from 1.0 mm to 5.0 mm depending on degree of agitation) and is therefore classified as a superflocculent yeast. Images of the various yeast flocs are presented herein.
Sampling Protocol
As the fermentations progressed, it was necessary to withdraw samples from the fermenting liquid at numerous time intervals. To perform this task, sterile sampling valves were purchased from SCANDI-BREW®. These valves are constructed of stainless steel and are equipped with a chamber (delimited by a top and bottom port) in which ethanol can be stored to maintain an aseptic environment. Before taking a sample, the ethanol is released from the chamber by removing the retaining cap from the bottom spout. Fresh ethanol (75 percent by volume) is run through the chamber, and the cap is then placed on the top port of the valve. The valve lever is then pulled, and a liquid sample of approximately 50 mL is collected into a sterilized container. A second sample is collected for the fermentations involving superflocculent yeast so that proper deflocculation can be performed prior to cell enumeration. Once the sampling is complete, the valve chamber is rinsed with hot water and peracetic acid and then, finally, with ethanol. The retaining cap is placed on the bottom spout and the chamber is filled with ethanol in preparation for the next sampling.
Microbiological Monitoring
Free Yeast Cell Enumeration and Viability by Methylene Blue Method
Liquid samples containing freely suspended yeast cells are first collected from the fermentation medium by the above sampling procedure. A Hauser Scientific Company Hemacytometer with a volume of 10−4 mL is used in conjunction with a light microscope to perform the cell counts. The liquid samples should be diluted with distilled water in order to achieve a total yeast count of 150 to 200 cells in the counting field. Heggart et al. (1999) describe all the factors that affect viability and vitality characteristics of yeast.
In order to assess the degree of viability within the sample, the methylene blue staining technique described by the American Society of Brewing Chemists was used (Technical Committee and Editorial Committee of the ASBC, 1992). Live cells can render the methylene blue stain colorless by oxidizing it. Dead cells, on the other hand, will stain blue. The following reagents were used in the preparation of methylene blue for viability assessment:
The Fink-Kuhles buffered methylene blue was then prepared by mixing 500 mL of solution A with 498.75 mL of solution B and 1.25 mL of solution C to yield a final mixture at a pH of 4.6.
A mixture of diluted cell suspension and methylene blue was prepared in a test tube and then thoroughly mixed. After allowing this mixture to rest for several minutes (ensures contact between cells and the dye), a drop of liquid was placed between the hemacytometer's counting glass and the cover slip (defined volume). The percentage of viable cells was determined by counting both the viable and dead cells within the counting field and then dividing the number of viable cells by the total number of cells.
Immobilized Yeast Cell Counts—Self Aggregation
When using yeast cells with a tendency to form flocs, it becomes difficult to accurately assess the number of cells present in a liquid sample because the cells will tend to settle in the sample jar. In order to obtain a representative sample, a deflocculating agent was used. In these experiments, a 0.5 percent by volume sulfuric acid solution was employed to destabilize the flocculated yeast cells, hence allowing for a representative yeast cell count. The same enumeration and viability procedure outlined below was used with the sulfuric acid replacing distilled water as the diluting agent.
Immobilized Yeast Cell Counts—Gel Beads
Before yeast counts were performed on gel-entrapped cells, it was necessary to disrupt the gel matrix using a POLYTRON® apparatus (Brinkmann Instruments). A sample of beads was first passed through a sterile sieve (500 μm mesh size) and then flushed with sterile water. One milliliter of gel-entrapped-cell beads and 19 mL of distilled water were added into a 50 mL sample container. The POLYTRON® was then used to disrupt the gel physically and thus release the yeast into solution. The enumeration and viability methods described in the following sections were then performed on the gel-disrupted sample.
Contamination Monitoring
All the fermentations performed throughout this invention were regularly monitored for contamination. The monitoring program consisted of at least one check per week of the liquid in the 50-L continuous fermenters and the wort in the storage vessels. Liquid samples were withdrawn aseptically and then spread onto culture plates composed of Universal Beer Agar (UBA, Difco Laboratories) and 10 mg/L of cycloheximide. These test samples were then incubated at 28° C. for up to 10 days in both aerobic and anaerobic conditions. Placing the selected plates into jars containing an ANAEROGEN® packet (Oxoid), which removes any oxygen remaining in the jar, created the desired anaerobic growth environment. The use of an indicator strip (turns pink if oxygen is present) allowed the inventors to verify that the environment was, indeed, anaerobic. Bacterial contaminants, if present in the liquid sample, would then be detected by this method.
Wild or non-brewing yeast detection required a separate growth medium that would not favor bacterial and/or brewing yeast growth. Pour plates prepared with yeast medium (YM, Difco Laboratories) supplemented with 0.4 g/L CuSO4 were utilized to selectively allow for the growth of any potential wild yeast (incubation at 25° C. for seven days). Incubating the liquid sample plated on PYN agar (Peptone Yeast-Extract Nutrient, Difco Laboratories) for seven days at 37° C. allowed for the detection of non-lager brewing yeast. Lager yeast growth is inhibited at temperatures above 34° C. Thus any growth on these plates would indicate an ale yeast contamination.
Analytical Methods
Appropriate calibrations were performed on all the relevant equipment as prescribed by standard industrial operating procedures.
Ethanol
Ethanol concentration in beer and fermenting samples was analyzed using the gas chromatography method described by the Technical Committee and the Editorial Committee of the American Society of Brewing Chemists (1992). A de-gassed sample of liquid was combined with five percent v/v isopropanol internal standard followed by the injection of 0.2 μL of this mixture into a Perkin Elmer 8500 Gas Chromatograph. The following list provides further detail regarding the exact setup of the GC:
Flame ionization detector (FID)
Dynatech autosampler
Chromosorb 102, 80-100 mesh support packing
Helium carrier gas flowing at 20 mL/min
Injector temperature of 175° C., detector temperature of 250° C. and column temperature of 185° C. isothermally.
Carbohydrates
The glucose, fructose, maltose, maltotriose, maltotetrose, polysaccharides and glycerol concentrations were measured using a high-performance liquid chromatography (Spectra-Physics SP8100XR HPLC) system. A cation exchange column (Bio-Rad Aminex HPX-87K) with potassium phosphate dibasic as the mobile phase was used to separate these carbohydrates as they eluted through the system. The quantity of the compounds was then determined using a refractive index detector to generate the appropriate compound peaks. The HPLC was operated at a back pressure of 800 psi, a column temperature of 85° C. and a detector temperature of 40° C. The samples were degassed and diluted to the appropriate levels. A 10 μL injection was then introduced into the system at a flow rate of 0.6 mL/min.
The brewing industry commonly uses another measure to assess the liquid's overall carbohydrate level. The liquid specific gravity expressed in degrees Plato was measured using an Anton Paar DMA-58 Densitometer. Filtered and de-gassed samples were transferred into a special glass u-tube, which was then subjected to an electronic oscillation. The frequency of the oscillation through the liquid was measured and then correlated to a liquid specific gravity (g/100 g or ° P). It should be noted that this measurement is an approximation of the sample's total carbohydrate concentration (or specific gravity) since the calibrations are performed on aqueous sucrose solutions at 20° C. whose specific gravity is the same as the wort in question.
Vicinal Diketones
Total diacetyl (2,3-butanedione) and total 2,3-pentanedione concentrations were determined using a Perking Elmer 8310 Gas Chromatograph equipped with an electron capture detector. A five percent methane in argon carrier gas flowing at 1.0 mL/min was used as the carrier gas, and the sample was passed through a J & W DB-Wax column. The injector temperature was maintained at 105° C. while the detector temperature was set at 120° C. A Hewlett Packard 7694E headspace autosampler facilitated the analysis. Quantification was calculated by evaluating the peak area of the selected sample component and then cross-referencing it to the 2,3-hexanedione internal standard calibration value.
In order to assess the “total” concentration of these compounds, it was first necessary to equilibrate these samples to 65° C. and then hold then for 30 minutes at this temperature. This pre-analysis sample handling allowed for the conversion of α-acetolactate and α-hydroxybutyrate into their respective diketones, diacetyl and 2,3-butanedione.
Esters and Higher Alcohols
Some of the most important flavor compounds detected in beer were measured using a headspace gas chromatography method. Acetaldehyde, ethyl acetate, isobutanol, 1-propanol, isoamyl acetate, isoamyl alcohol, ethyl hexanoate and ethyl octanoate were quantified using n-butanol as the internal standard. A Hewlett Packard 5890 Gas Chromatograph equipped with a flame ionization detector, a HP 7994 headspace autosampler, and a J & W DB-Wax capillary column was utilized. The injector temperature was set at 200° C., and the detector temperature was 220° C. The oven temperature profile was as follows: 40° C. for five minutes, ramp from 40° C. to 200° C. at a rate of 10° C./min, ramp from 200° C. to 220° C. at a rate of 50° C./min, and finally a hold at 220° C. for five minutes. A helium makeup gas at 30 mL/min (28 psig), a hydrogen stream at 50 mL/min (25 psig), and an air stream at 300 mL/min (35 psig) supplemented a helium carrier gas flow of 6.0 mL/min. The entire GC cycle for a sample loop of one mL was 40 minutes.
Other Analyses
Several other analytical measurements were performed on an as-needed-basis on fermentation liquid that was subjected to aging and packaging. Finished product analyses were performed by the Labatt Quality Control department as per finished beer standards. The methods described by the Technical Committee and the Editorial Committee of the American Society of Brewing Chemists (1992) were the basis for these measurements. A list of the analyses, as well as a brief description of the relevance of these measurements, is provided in Table 2.
Yeast Settling Protocol—LCC290
Preparation of Yeast Test Sample
Superflocculent yeast (LCC290) was grown in wort previously described. This inoculum was then centrifuged at 4° C. and 10000 rpm for 15 minutes in order to obtain a yeast pellet for further inoculation. Wort was pasteurized at 100° C. for 60 minutes, and then one liter was aseptically transferred into 6×2 L sterilized shake flasks. Each shake flask was inoculated with four grams of centrifuged yeast. The flasks were placed on a shaker operating at 135 rpm (21° C. ambient room temperature) and allowed to ferment. One flask was withdrawn at the following time intervals: 24 h, 40 h, 48 h, 64 h, 71 h, and 192 h. At each interval, a small sample of liquid was taken for carbohydrate analysis and for yeast concentration and viability measurements. The remaining liquid/yeast mixture was subjected to the yeast settling protocol described below.
Yeast Settling Protocol
The settling rate for LCC290 superflocculent yeast strains was measured using the following method. Each sample was allowed to ferment as described above. At the prescribed time, the appropriate sample flask was withdrawn. The samples were agitated in order to ensure that all particles were suspended. The contents of the flask were then immediately transferred into a 1000 mL graduated cylinder. As the flocs settled, the distance between the liquid surface and the floc-liquid interface was measured at 30 second intervals. The settling rate was calculated by applying the following equation:
Using the standard Kynch method (1952), a settling rate versus cell concentration curve was generated from the settling curves obtained for each fermentation interval.
Circulation and Mixing Rate Methods
In order to measure mixing time and circulation rate inside the three phase gas-lift draft tube bioreactor, an acid injection system linked to a data acquisition system was utilized. By applying a pulse of a strong acid into the bioreactor, it was possible to calculate both circulation rate and mixing time by monitoring the change in pH over time and then relating them to the equations presented in section 3.2.2. The data acquisition system consisted of:
an Ingold pH probe (Cole-Parmer, cat. #P-05990-90) coupled to an Ingold Microprocessor-Based pH Transmitter (Model 2300)
a Data Translation DT2805 card
a 386DX personal computer and
a Quick Basic data acquisition program (written by C. Hudson and J. Beltrano in 1994 and modified by N. Mensour in 1998).
Ten milliliters of 10 N hydrochloric acid were injected into the annulus section of the gas-lift draft tube bioreactor (diagram is provided in
The collected data were transferred from the laboratory to a more powerful computer (Pentium II microprocessor) for further analysis. TableCurve 2D (Jandel Scientific Software, Labtronics, Canada) was utilized extensively for data analysis due to its capacity to deal with large data sets, as well as its many built-in data handling functions (data smoothing, curve fitting, etc.). The Savitzky-Golay algorithm, a time-domain smoothing method based on least squares quartic polynomial fitting across a moving window, was applied to the original data to eliminate noise. The smoothed data were then adjusted to reflect a change in pH rather than the actual pH measurement. A decaying sinusoidal function was fitted to the adjusted data. The mixing times and circulation rates were then calculated from the fitting parameters.
These mixing experiments were conducted on actual fermentations inside the 50-L gas-lift bioreactor with one of three immobilization carriers present (either superflocculent yeast LCC290, medium flocculent yeast LCC3021 or κ-carrageenan gel beads). The liquid phase was fermented beer with a specific gravity of 2.5° P, and the gas phase was comprised of carbon dioxide sparging gas. The fermentation temperature was controlled at 15° C. The sparging gas superficial velocities were varied between 2.0 and 6.0 mm/s. At a given gas flowrate, the system was allowed to equilibrate for 10 minutes. The acid injection test would then start and be repeated three times. The next gas flowrate would be selected, and the mixture inside the reactor would have its pH readjusted to the starting level.
Design of a Pilot Scale Gas-Lift Draft Tube Bioreactor System
Gas-Lift Draft Tube Bioreactor Fermentation System
Draft tube fluidized bed (DTFB) systems have shown their value for use in three phase systems. Two identical pilot scale gas-lift draft tube bioreactors were designed, built and installed in the Experimental Brewery of the Labatt Brewing Company Limited in order to carry out the experimental work for this invention. In addition, several existing vessels were modified for both wort storage and beer collection. The flowsheet in
Wort was supplied from the London Brewing plant through 5.08 cm stainless steel lines and transferred into 1600 L working volume wort storage tanks (WT1 & WT2). With a two-tank system, it was possible to ensure a continuous supply of nutrient medium to the pilot scale continuous fermenters (R1 & R2). Each holding tank is equipped with a carbon dioxide sparge system for oxygen and homogeneity control and a glycol cooling jacket system for temperature control. This central source of nutrient medium was set up to feed up to three independent fermenters through a valve header system (V7, V8 & V9). Masterflex peristaltic pumps (P1 & P2) were utilized to deliver a prescribed flow of wort to the pilot scale bioreactors (R1 & R2).
Carbon dioxide gas and air were sparged into the gas-lift system through stainless steel gas pipe spargers (
More detailed diagrams and exact dimensions of the 50-L pilot scale bioreactor are provided in the following sections, including wort handling and storage protocol and cleaning and sterilization protocols for the continuous fermentation system.
Reactor Design and Specifications
The 50-L working volume bioreactor designed for this work was built entirely of 304 L stainless steel with four Plexiglas look windows located in the body of the reactor so that particle and fluid motion could be observed. The material of construction was chosen for its resistance to sanitation chemicals (caustic and acid), as well as for its durability to steam sterilization. Another important aspect of the design was the minimization of threaded fittings in direct contact with the fermentation medium. Instead, ports were welded and, where necessary, sanitary TriClover fittings were used. The reactor was designed with an expanded head region so as to maximize gas disengagement and thus promote better liquid-solid mass transfer (Chisti & Moo-Young, 1993). The reactor bottom was designed with a 90-degree cone angle so as to minimize any solids from collecting at the bottom.
The internal draft tube and the particle separator (baffle) are illustrated in
A pipe sparger (
Wort Handling and Storage Protocol
In traditional fermentation practices, wort is not held for extended periods of time without being pitched with yeast. Oxygenated wort is an excellent growth medium for many organisms, including yeast. Because the fermentation protocol for the continuous gas-lift systems required large quantities of wort to be held, it was necessary to develop wort transfer and holding protocols. It was the opinion of the researchers that unoxygenated cold wort could be stored for up to two weeks without it being compromised by contamination, if the wort was transferred to the holding vessel appropriately.
It was also deemed necessary to ensure that the temperature of the wort was controlled appropriately once inside the wort holding vessel. The available tankage inside the Experimental Brewery was originally designed for fermentation rather than wort holding. A test of these tanks' capability to maintain the temperature of stagnant liquid was performed.
By introducing slight agitation through the injection of carbon dioxide at a flow rate of 0.133 cm3/hour, it was possible to maintain the temperature of the wort inside the holding vessel at 4° C. As a result of these findings, both wort holding vessels were fitted at the base of the cone with a 2.54 cm sanitary tubing to be used for wort agitation.
Unaerated wort was subsequently transferred from the Labatt London plant through the 5.08 cm stainless steel line into a buffer tank. From this tank, the wort was passed through a flash pasteurizer into one of the wort holding tanks (WT1 or WT2) where it was stored at 2° C. for up to two weeks. This pasteurization step was put in place as a precautionary measure to ensure that unwanted microorganisms were eliminated from the wort during the entire holding period. By utilizing unoxygenated wort, the damage to the hot wort by oxygen (formation of staling aldehydes) would be minimized. In addition, the air introduced with the sparge gas could therefore strictly accomplish the control of oxygen to the continuous fermenters.
Dissolved oxygen measurements were performed on the wort once inside the wort holding vessel.
In the final trial, the tank was pre-purged as above, and a continuous carbon dioxide sparge (0.085 m3/h) was introduced through the filling process, as well as during the hold period. The dissolved oxygen content was kept to a minimum (<0.1 mg/L) throughout the holding phase. Consequently, this method was adopted as the protocol for all future wort collections.
Cleaning and Sterilization Protocol
The wort holding tanks were subjected to a cleaning cycle consisting of a pre-rinse with hot water (85° C.), a caustic cleaning rinse (40 percent caustic at 60° C.) followed by a hot water post-rinse (85° C.). Sanitization of vessels was accomplished by contacting the walls with a peracetic acid solution (two percent w/v). The piping wort transfer piping also went through the same cleaning and sanitization regiment.
The 50-L bioreactors followed a different cleaning and sterilization protocol. The systems were rinsed with hot water (60° C.) and then filled to the top with 40° C. warm water. An industrial cleaning agent, Diversol CX/A (DiverseyLever, Canada), was then added to this water to form a two percent w/v solution. Air was sparged into the bottom of the reactor at a superficial gas velocity of five mm/s to ensure proper dissolution and proper contacting within the reactor. After one hour contact time, the reactor was emptied and flushed with fresh city water. This cleaning procedure was repeated a second time, culminating in two final cold city water fill-empty cycles.
The waste beer vessel was cleaned using a two percent w/v Diversol CX/A solution. Unlike the gas-lift bioreactors, sparging was not utilized since the WBT was not fitted with a sparger. Mechanical agitation was accomplished by rolling this vessel on its side. The cycle was repeated twice and was followed by two water fill-empty cycles.
Prior to steam sterilization, the 50-L bioreactors were connected to the waste beer vessel and the wort feed line was disconnected, as was the gas sparge line. See
The wort feed line was connected to the bioreactor after an internal temperature of 20° C. was reached. With valves V2, V5, V10 and V14 still in the closed position and with valves V6, V7, V8, V9, V11, and V15 open, the steam supply was connected at either V3 or V6, depending on the wort supply being utilized. The steaming cycle lasted for one hour after which time valves V9, V11, and V15 were shut off simultaneously with the steam supply. Once the lines had reached room temperature (20° C.), valves V3 and V6 were closed, and the steam supply was disconnected. At this point, the entire continuous fermentation system, including the wort supply, the 50-L bioreactors and the waste beer tank, was sterilized and ready for fermentation.
Fermentation Protocol
The 50-L pilot scale gas-lift draft tube bioreactors were used for the continuous primary fermentation of brewer's wort into beer. A glycol thermal jacket provided temperature control with a liquid temperature of 15° C. targeted throughout the fermentation trials. Each reactor was equipped with a temperature probe for measurement purposes and a temperature thermocouple and glycol solenoid valve for the adjustment of glycol feed to the reactor. The gas-lift fermenters were also equipped with a primary mixing gas (carbon dioxide or nitrogen), as well as with an air supply for oxygen dosing. The desired mixture of gas was selected by adjusting the appropriate rotameter/needle valve combination and then passing this gas mixture through the sterile filter (Millipore, MILLEX®-FG50, 0.2 μm Filter Unit) and into the draft-tube of the bioreactor. A superficial air velocity of 0.39 mm/s (0.4 scfi) was injected into the reactor for all the fermentations, while the primary mixing gas flowrate was adjusted to suit the specific immobilization type.
The 50-L gas-lift bioreactor followed a traditional batch start-up before a continuous mode of operation was started. After cleaning and sterilization as described above, the gas-lift bioreactor was filled with 50 liters of wort from the wort holding tanks (WT1 or WT2) and then injected with 200 grams of yeast (4 g/L) through the SCANDI-BREW® sterile sample port. In the case of the k-carrageenan gel beads, 20 L of beads were injected into the reactor, yielding an initial concentration of LCC3021 medium flocculent yeast of four grams per liter. The bioreactors were sampled daily, and the evolution of diacetyl and the liquid specific gravity were closely monitored. Once the specific gravity had reached its minimum value and the diacetyl concentration had dropped below 30 μg/L, it was deemed that the system could be set into continuous operation.
The fermentation medium (wort) was continuously fed through the bottom of the reactor while “green” beer overflowed through the funnel at the top of the reactor. As the reactor's working volume was fixed, selecting the flowrate of the fresh wort feed into the reactor controlled the average liquid residence time. Liquid samples were withdrawn from the reactor daily at the outlet through the sterile sampling valve (SCANDI-BREW®) for both chemical and microbiological analyses.
At selected time periods, continuous fermentation product was collected from the 50-L primary fermentation bioreactor in larger quantities (40-L sterile stainless steel cans) and subjected to post-fermentation processing in order to produce a finished, saleable beer for evaluation and comparison to industrially-produced control beer. The selected 50-L bioreactor was disconnected from the waste beer vessel and immediately connected to the beer collection vessel. Once the desired liquid had been collected, the bioreactor was reconnected to the waste beer vessel. The collected “green” beer was subjected to a post-fermentation hold period in order to reduce the liquid's diacetyl level below 30 μg/L. The yeast carried over with the liquid was allowed to settle, and the liquid (cells concentration of ˜1-5 million cells/mL) was placed in cold storage for aging (seven days at 2° C.). After the aging period, the liquid was filtered, diluted to five percent alcohol by volume, and carbonated before being packaged in 341-mL beer bottles. All packaged liquid was then subjected to pasteurization through Labatt plant equipment.
Continuous Gel Bead Production Process
The objective of this section of experimental work was to evaluate a continuous bead production process for the production of yeast-inoculated gel beads in order to supply immobilized LCC3021 yeast cells to the 50-L continuous gas-lift draft tube bioreactors previously described.
The production process (
The formation of the κ-carrageenan gel bead emulsion was conducted in a 37° C. temperature controlled water bath in order to prevent premature gelification of the carrageenan gel. The sterilized polymer was maintained at 37° C. in a temperature regulated water bath and the yeast inoculum was maintained at 20° C. prior to immobilization. Using Masterflex peristaltic pumps (Cole Parmer Company, USA), the gel and the yeast slurry were pumped through 24 elements of the 6.4 mm diameter static mixer in order to disperse the cells evenly through the gel. The sterilized oil, stored at room temperature, was pumped (Masterflex peristaltic pump) into the hot water bath to also reach a temperature of 37° C.
The inoculated polymer (aqueous phase) was then mixed with the oil (continuous phase) through another series of static mixers to create the desired emulsion. This resulting emulsion was rapidly cooled to 5° C. inside a water/ice bath, provoking the gelling of the polymer droplets into beads. The beads then proceeded into a sterile 22 g/L potassium chloride solution which aided their hardening and allowed for their separation from the oil phase. The process oil was recycled back to the process and the aqueous phase (beads and potassium chloride solution) was transferred into a separate tank for size classification before loading into the 50-L bioreactors.
Static Mixer—Kenics Type
At the heart of this novel bead production process are Kenics static mixers (Cole Parmer Instrument Company, Niles, Ill., USA). They are composed of a series of stationary elements placed in a tube with an internal diameter equivalent to that of the static mixer diameter. These elements form crossed channels, which promote the division and the longitudinal recombination of the liquid flowing through the static mixer. The transversal rupture of these finely created streamlines into an increasingly homogenous emulsion is furthermore provoked by this mixing system. Table 4 lists the three types of static mixers that were used in this study.
Gel Production Materials
The two principal materials in the production of the gel beads were the oil and polymer. κ-Carrageenan (type X-0909, lot 330360, Copenhagen Pectin, Denmark), a polysaccharide polymer extracted from red algae, was a generous gift from Copenhagen Pectin A/S. This polymer possesses the unique property of thermo-gelation, where its gelling temperature depends on the concentrations of both κ-carrageenan ([Car]) and potassium chloride ([KCl]). The polymer was dissolved to a concentration of 30 g/L into distilled water at 80° C. containing 2.0 g/L of KCl. The resulting gel solution had a gelling temperature of 28° C. The gel was autoclaved for one hour at 121° C. and then placed into a 40° C. water bath so that it would not harden. Commercial grade corn oil (Pasquale Bros. Inc., Canada) was also sterilized for one hour at 121° C. and then stored at room temperature (20° C.) until its use. A yeast slurry was prepared as described in section 4.1.
Measurement of Bead Diameter
Bead samples were collected at the exit of the 5° C. heat exchanger in flasks containing 100 mL of 22 g/L KCl solution. The beads were allowed to soak in this solution for two hours to promote their hardening. The oil was removed from the aqueous phase by successive washes with potassium chloride solution. The samples were then stored at 4° C. to prevent microbial contamination prior to analysis.
The measurement of bead diameter was performed using the image analysis software Optimas (Version 4.02, BioScan, Inc, USA) linked to a video camera (Pentax Macro 50 mm). A bead sample was transferred into a petri dish containing a fine film of water (used to separate the beads) and then placed under the camera. A total of 300 to 400 beads were measured using this system. The capabilities of the Optimas software lie between 100 μm and several mm with a maximum absolute error of 30 μm. The data obtained from Optimas were analyzed further using Microsoft Excel. The resulting sample size distributions were characterized by their sample mean diameter (DB) and coefficient of variation (COV).
Bead System Evaluation—Experimental Plan
A total of three static mixer diameters (Ds=6.4 mm, 9.5 mm and 12.7 mm) were compared to assess which type of bead population would be produced as measured by the sample's average bead diameter and coefficient of variation of the size distribution. The number of static mixer elements (Ns) was varied between 12 and 120 elements while the polymer volume fraction (εc) was studied between 8.3 percent v/v and 50 percent v/v gel in oil solution. Above an εc of 50 percent, the dispersed (gel) and continuous (oil) phases became inverted, that is, oil droplet inclusion within the polymer matrix resulted instead of gel droplets within the oil matrix. The superficial liquid velocity of the oil/gel emulsion through the emulsion section was adjusted in the range of 3.6 cm/s and 17.8 cm/s. The superficial liquid velocity (VSL) through the emulsion static mixer was calculated by the following equation:
VSL=(Qoil+Qcar)/S
where S is the cross sectional area of the tubing which contains the static mixer, Qoil is the volumetric flowrate of the oil phase, and Qcar is the volumetric flowrate of the carrageenan gel solution.
Results and Discussion of Fermentations and Mixing Dynamics within 50-L Gas-Lift Bioreactors
The use of immobilized cells for the production of ethanol has been published. In the last two decades, researchers have attempted to optimize the ethanol production process by coupling immobilized cell technology with continuous processing (Kuu, 1982; Gil, 1991; Maiorella, 1983). Many have met with great success and the use of continuous immobilized cell systems for ethanol production has become industrial. However, the implementation of such a continuous process in the brewing industry for the primary fermentation of beer is not as simple. Beer is comprised not only of ethanol but also of a myriad of flavor compounds, which add both complexity and depth to the final product. The following chapter describes the results obtained by the inventors in the quest for producing a well-balanced beer within the pilot scale GLDT fermenter.
Batch Fermentations in the Pilot Scale GLDT System
Batch fermentations utilizing freely suspended yeast cells were conducted in the 50-L pilot scale gas-lift draft tube bioreactor. These trials provided the opportunity to assess the feasibility of using such a system for the fermentation of wort into beer. In addition, the trials served to establish a benchmark for future comparison with continuous fermentation liquids. Two batch fermentations were undertaken in the 50-L bioreactor using a lager yeast strain from the Labatt Culture Collection (LCC3021). The yeast growth rate, as well as the consumption of nutrients and release of products, was monitored throughout the fermentations.
Ethanol is one of the most important by-products of yeast metabolism. An optimized anaerobic fermentation will produce about 48 g of ethanol and 47 g of carbon dioxide per 100 g of metabolized glucose. Small quantities of glycerol will also be produced (3.3 g per 100 g glucose) as this by-product is involved in maintaining the redox balance within the fermenting yeast, as well as supporting the cell in its osmotic balance, particularly in hypertonic media.
From the results presented in this section, it appeared that the two batch fermentations proceeded normally within the gas-lift draft tube bioreactor. Yeast growth and carbohydrate uptake followed the expected paths, as did the by-products, ethanol, diacetyl and pentanedione. A comparison of the individual batch data indicates that the two batches fermented in very similar fashion.
Immobilization Carriers
Several carriers were investigated throughout this invention to identify the most promising alternatives for future development work. Three distinct modes of immobilization were tested within the 50-L gas-lift draft tube bioreactor. Two commercially available adsorption carriers with sizes ranging between one and two mm were evaluated. SIRAN®, a glass bead carrier supplied by Schott Engineering (
The initial fermentation results with both of these carriers in the 50-L gas-lift draft tube bioreactor were unfavorable. The problems that arose were mainly due to the high particle densities of SIRAN® and CELITE® as compared to the liquid medium. Three phase gas-lift draft tube systems work best when the ratio of the carrier and the liquid densities is kept close to unity. In the case of SIRAN®, the ratio was 1.34, whereas the ratio for CELITE® was 1.31. The consequence of having such high density differences between the solid and liquid phase was a significant increase in the minimum gas fluidization velocity required to operate the 50-L gas-lift draft tube bioreactor. For four liters of SIRAN® carrier (eight percent v/v solids loading), a gas velocity of 21.5 mm/s (based on draft tube diameter) was required in order to achieve circulation. This higher gas velocity was not a significant problem when testing was performed in a water solution. However, as soon as the liquid medium was wort, catastrophic failure occurred within the gas-lift draft tube (GLDT) system. The required gas velocity caused excessive foaming within the reactor, which ultimately reduced the liquid level to below the draft tube, effectively stopping liquid and solid circulation. Failure similar to that encountered with the SIRAN® occurred when CELITE® was substituted as the immobilization material. Due to these results, both of these adsorption-based carriers were abandoned in future gas-lift fermentation trials.
Using an entrapment-based carrier like κ-carrageenan allowed the system to be loaded on a 40 percent (v/v) basis with solid and required about 0.17 standard cubic meters per hour of gas (5.8 mm/s superficial gas velocity) for its fluidization and subsequent circulation. The positive results experienced with operating the system with carrageenan-entrapped yeast cell beads were due to the carrier's lower density (about 1100 kg/m3) and consequent ease of fluidization. Similarly, desired solids loading with the self-aggregating yeast, LCC 3021 (medium flocculent) and LCC290 (super flocculent), were achieved with gas fluidization velocities of approximately three mm/s required to ensure proper circulation. The following sections describe in more detail the k-carrageenan gel carrier and describe the self-aggregating yeast—LCC3021 flocs and LCC290 flocs which were evaluated as immobilization matrices for continuous primary fermentation within the 50-L GLDT fermenter.
κ-Carrageenan Gel Beads
Entrapment-based immobilization methods require the inclusion of the yeast cells within the matrix prior to their introduction into the fermentation vessel. Since in-situ reactor inoculation is not feasible at this time, it is necessary to produce these gel beads prior to the commencement of fermentation. It is still not clear what effects long term storage has on inoculated gel beads. In order to minimize any potential negative storage effects, it was decided to produce large quantities of gel beads within a short period of time (eight hours). The static mixer bead process described above was therefore utilized for this purpose. The ideal beads would have particle diameters (DB) between 0.8 mm and 1.4 mm with the coefficient of variation (COV) of the size distribution kept to a minimum. It was necessary to adjust several parameters of the bead making process to produce the desired amount and consistency of beads. The following sections present a summary of the bead process parameter selection and describe the beads used in the continuous fermentation trials.
Bead Production Process: Variable Selection
Characterization of the bead making process was undertaken in collaboration with other researchers with emphasis placed on the following process parameters static mixer diameter (Ds), number of static mixer elements (Ns), superficial liquid flowrate (VSL) and polymer volume fraction (εC). FIGS. 22 to 31 summarize the results obtained through experimentation.
All of the data collected from this bead making process showed size distributions with only one peak. Poncelet et al. (1992), however, showed the occurrence of satellite peaks and/or a secondary peak corresponding to beads with diameters smaller than 200 μm for alginate beads produced by dispersion in a stirred tank. It is possible that smaller beads produced in the inventors' process were simply lost during the bead-washing step and therefore would not appear in the inventors' size distribution data.
The effects of superficial liquid velocity and static mixer diameter on the average bead diameter and on the coefficient of variation of the size distribution are depicted in
At a superficial liquid velocity of 3.5 cm/s, the polymer volume fraction appeared to affect the average bead diameter, while velocities above 7 cm/s produced relatively no differences at experimental values of εc varying between 0.083 and 0.5 (
It had been hypothesized that an increase in the static mixer diameter (Ds) would create heterogeneity of shear forces within the mixer, inducing an increase in the size dispersion as measured by the coefficient of variation. Concurrently, an increase in Ds would decrease the intensity of the shear forces thus increasing the mean bead diameter. Both of these effects were observed in the experimentation with the smallest diameter static mixer producing beads with the smallest average bead diameter (400 μm-500 μm) and the smallest coefficient of variation (approximately 40 percent) or size dispersion.
The energy required to create an emulsion is proportional to the interfacial area created by the polymer and the oil phase. The smaller the bead size, the larger the energy required for formation. Berkman and Calabrese (1988) have shown that an increase in the average superficial liquid velocity (Vs) provokes an increase in the dissipated energy per unit mass of fluid, thus favoring a reduction in the bead size. An increase in the average superficial liquid velocity (tested between 3.6 cm/s and 17.8 cm/s) produced a decrease in the average bead size. Such a velocity increase results in a pressure differential between the static mixer inlet and outlet. This pressure differential is proportional to the dissipated energy per unit mass of liquid. An increase in velocity therefore induces an increase in the system's dissipated energy, which favors a reduction in bead size. The reduction in bead diameter, DB, was observed as the superficial liquid velocities increased. Al Taweel and Walker (1983) have shown that a dynamic equilibrium is established between the formation of beads and the coalescence between beads for high velocities corresponding to significant turbulence levels. For constant static mixer diameter (Ds) and number of elements (Ns), the superficial velocity had little effect on the coefficient of variation. Velocity is therefore a parameter, which allows the manipulation and selection of the average bead diameter without significantly modifying the size dispersion.
Within the scope of this research, the carrageenan gel volume fraction (εc) had little effect on either the average bead diameter or the coefficient of variation, except for the lowest studied velocity of 3.6 cm/s where the average bead diameter decreased with a decrease in εc. Audet and Lacroix (1989) studied this parameter for the production of carrageenan beads in a two-phase dispersion system (batch stirred tank not continuous static mixer process) and they concluded that εc had no effect on the mean bead diameter for a polymer solution with a carrageenan concentration of 3 percent (w/v). The specific effect of the κ-carrageenan gel concentration on the bead size distribution was examined by Audet and Lacroix (1989) who showed that this parameter strongly influenced the size distribution. Increasing gel concentrations resulted in increasing average bead diameter (DB) and coefficient of variation (COV). The noted effect was attributed to the increased viscosity of the gel at higher concentrations resulting in lower shear forces on the emulsion and thus larger beads. Although the effect of gel concentration on bead size was not investigated in this invention, it could be used as another means of controlling bead size if necessary.
An increase in the number of mixing elements (Ns) increases the average residence time that a fluid element spends inside the static mixer, resulting in a more homogenous mixture and thus the formation of smaller and more tightly dispersed beads. In the experimentation, an equilibrium in the dispersion (measured by the coefficient of variation) was reached around 60 to 72 elements. Middleman (1974) has shown that 10 elements were sufficient to attain such an equilibrium in the case of emulsions with low viscosity (0.6 to 1.0 cP). The carrageenan solution used in these experiments [three percent w/v)] had an average viscosity of 200 cP and the oil's viscosity was 25 cP. Consequently, this higher viscosity required a longer residence time inside the mixer in order to reach pseudo homogeneity.
Bead Production Process: κ-Carrageenan Bead Characteristics
From the data described in the previous section, it was possible to select bead production process parameters to produce beads with the desired characteristics for the fermentation trials. To minimize the coefficient of variation for a particular static mixer diameter, 60 mixing elements were chosen to create the oil-gel dispersion. Beads with an average diameter of approximately one mm were selected to minimize external mass transfer and to facilitate separation by mechanical means from the fermenting liquid. To achieve this, the largest static mixer tested (12.7 mm) and the lowest tested velocity (3.6 cm/s) were selected. Since the polymer fraction had little effect on the coefficient of variation, a 50/50 ratio (ε=0.5) of gel to oil was employed in order to maximize bead productivity.
Several batches of gel beads, with LCC3021 yeast entrapped, were produced in the laboratory using the process described previously with Ds=12.7 mm, Ns=60, VSL=3.9 cm/s and εC=0.5. The resultant beads (
Limitations of the Bead Process with Respect to Industrial Scale Application
A process producing 10 L of beads per hour, per static mixer, was developed and implemented at a pilot plant level. Several aspects of this process require further development and/or optimization before industrial scale bead production can be considered. An increase in volumetric productivity of the system is necessary in order to supply the volume of immobilized cells required to feed a large-scale bioreactor. For example, a 2000-hL gas-lift draft tube bioreactor would require approximately 800 hL of beads. To achieve such volumes, an increase in both the flows of gel and oil are necessary. The data suggest that the resulting increased velocity using the static mixers of 6.4 mm to 12.7 mm diameters would induce the formation of beads too small to be used in the fermentation system. It would therefore be necessary to increase the diameter of the static mixers, thus increasing the average bead diameter. However, the use of static mixers with a larger diameter will also increase the bead size dispersion, producing a larger percentage of beads outside the desired range.
Another alternative would be the implementation of a system using static mixers of medium size (12.7 mm) placed in parallel. With a ten static mixer system, productivities reaching 100 L/h are conceivable. For the 2000-hL industrial example, the process would run continuously for 800 hours or approximately 34 days to produce the required volume of beads. Several systems could be implemented to reduce the production time, but this would add yet another level of complexity. The production time could become less of an issue if it were possible to store the beads for extended periods of time while retaining yeast viability. It is conceivable that a process for drying the beads or storing the beads in a vacuum-sealed container could be developed.
Poncelet and co-workers (1993) have published work indicating that the type of static mixer used to create the dispersion must be considered. With their proposed system using another type of static mixer, as opposed to the Kenics-type used in this research, it may be possible to utilize a larger diameter mixer without compromising the size distribution of the beads (maintaining a low coefficient of variation).
Additional concerns with the existing pilot process include operating the system at 40° C. and the use of vegetable oil and potassium chloride solutions for bead production. Due to the high production temperature, both heating and cooling systems are required in the process. The potential thermal shock that the yeast cells are exposed to requires additional investigation so that an assessment can be made as to potential negative implications. In this study, immobilized cell beads with high yeast viabilities (above 90 percent) were produced but any other effects that the process may have caused on the yeast population were not investigated.
Because oil is used to produce the desired emulsion and consequently the formation of the beads, and since oil will act as a surfactant thereby suppressing foam formation, the residue of oil on the bead surface is an issue. Although this residue would help during the fermentation stage, any carryover into the final beer would be detrimental, as foam is desirable in finished products. The large volumes of potassium chloride solution utilized to separate the solid phase (beads) from the oil and the method of removing this saline solution from the bead slurry before the introduction of beads into the bioreactor requires attention. Otherwise it may be necessary to flush this solution from the reactor following the addition of beads to the reactor.
Since the immobilized cell beads are produced outside the bioreactor, aseptic techniques must be utilized throughout the bead formation process and sterility maintained until the beads are introduced into the bioreactor. The various transfer points between tanks provide opportunity for contamination and must be monitored due to the fact that the presence of a contaminant could result in its co-immobilization within the bead. As a result of diligence within the laboratory, it was possible to consistently produce aseptic beads. However, the environment within a plant setting may not be as hospitable as the laboratory, therefore requiring much stricter control.
Flocculent Yeast Cells
One of the most natural forms of immobilization is the self-aggregation of microorganism into flocs of cells. Calleja and Johnson (1977) have proposed three possible reasons for cells to come in contact with each other to form aggregates, with all distinctive bonding properties. The first involves cells of different sexes being attracted to each other by the release of pheromones (α and α-factors). This type of bonding is temporary and involves protein-protein bonding between α and a-agglutinins anchored in the complementary cell walls.
Cells may also aggregate through their failure to separate from the mother cell during the budding process. This failure may be inherent to the particular yeast strain or can be caused by nutrient deprivation or mutation of a number of genes. This phenomenon is referred to as chain-formation and not flocculation. The bonds between these cells can be irreversibly destroyed by mechanical shear (Stratford, 1996).
The third scenario is more commonly known as flocculation. Stewart and Russell (1981) have defined flocculation as a reversible “phenomenon wherein yeast cells adhere in clumps and either sediment rapidly from the medium in which they are suspended or rise to the medium's surface.” Extensive evidence indicates that flocculation is genetically controlled and the mechanism of flocculation relies on selected molecules acting as bridges between adjacent cell walls. More specifically, it is thought that specific lectins are bound to the α-mannans of the adjoining cells in the presence of Ca2+ ions (Calleja and Johnson, 1977). This protein/carbohydrate bonding was found to be reversibly inhibited by chelating agents or by specific sugars.
Several measurement methods have been proposed for the assessment of yeast flocculation (Speers & Ritcey, 1995; Akiyama-Jibiki et al., 1997; Teixera et al, 1991; Stewart & Russell, 2000). In “Brewer's Yeast” (Stewart & Russell, 2000), it has been proposed that yeast flocculation methods can be subdivided into three categories, namely, sedimentation methods, static fermentation methods, and direct observation of floc formation in growth medium.
The sedimentation method first described by Bums in 1937 was modified by Helm and colleagues in 1953 and is currently part of the standard methods of analysis recognized by the Technical Committee and the Editorial Committee of the American Society of Brewing Chemists (1992). This technique is referred to as an in vitro technique as the yeast's settling characteristics are assessed in a calcium-sulfate buffer and not in the actual fermentation medium. The static fermentation method (also known as the Gilliland method) involves growing the yeast in hopped wort and measuring its flocculation characteristics in vivo. Both these methods involve measurements of the absorbance of settled yeast samples vs. yeast samples that have been deflocculated using a UV/Visible spectrophotometer.
Stewart and Russell (2000) present a measurement method for yeast flocculation by visually describing the level of flocculation that occurs in samples of yeast grown in 20 mL screw capped glass bottles. To express the degree of flocculation, they used a subjective measure, for example: 5—extremely flocculent, 4—very flocculent; 3—moderately flocculent, 2—weakly flocculent, 1—rough and 0—non-flocculent. The superflocculent yeast strain, LCC290 received a classification of 4—very flocculent whereas the LCC3021 yeast strain was classified as 3—moderately flocculent.
Flocculence is an important characteristic in the brewing industry as the yeast's natural tendency to either sediment or rise to the surface is commonly used as a separation method for this yeast from the fermenting liquid. However, a yeast strain that flocculates before the fermentation has completed is undesirable since the liquid will not have reached its ideal alcohol and residual sugar level. In continuous fermentation and in particular, gas-lift draft tube continuous fermentation, flocculent yeast act as the immobilization matrix. Their tendency to settle is compensated for by the injection of the sparging gas, which keeps them in suspension. With such a system, the fear of under-fermenting the liquid medium is eliminated since the solid particles are continuously circulated and kept in intimate contact with the fermenting liquid.
In the following section, the settling properties and the fermentation performance of the superflocculent yeast, LCC290, were characterized. The interest was in identifying the onset of flocculation for this particular yeast strain. In addition, the settling velocity of the yeast was determined in order to provide valuable information that could be used for the future in the design of yeast settling vessels.
Characterization of Superflocculent Yeast, LCC290
Before performing continuous fermentations with the superflocculent yeast LCC290 within the gas-lift draft tube bioreactor, it was decided to characterize the yeast in lab scale, shake flask fermentations.
From a growth and carbohydrate metabolism perspective, it appeared that the superflocculent yeast LCC290 behaved, as would the industrial yeast strain LCC3021. The fermentation appeared to reach completion when the liquid specific gravity reached approximately 2.7° P. It is common for flocculent yeast strains to form large clumps (flocs) and settle out of solution before end-fermenting. This phenomenon is known in the brewing industry as a ‘hung’ fermentation. In the inventors' batch trials, the inventors were able to end ferment because the flasks were agitated therefore keeping the yeast in suspension and in close contact with the nutrient supply.
Another important characteristic that was investigated was the yeast's ability to flocculate. In particular, the inventors were interested in establishing the speed, with which the yeast would settle, as well as get an indication of when this particular yeast strain would commence flocculation. Both of these characteristics are of importance for the continuous fermentation trials as they play a role in maintaining a healthy yeast population within the gas-lift fermenter.
Researchers have proposed that settling rate is a function of solids concentration (Coe and Clevenger, 1916).
The results collected during the settling test indicated that the superflocculent yeast strain LCC290 will flocculate at liquid specific gravities of 6° P and lower. This value can be used as a guide for continuous fermentations to indicate where the pseudo-steady-state liquid specific gravity should be kept if it is desired to keep the cells flocculated. Operating the reactor above 6° P would risk destabilizing the flocculated cells and possibly lead to the washout of the immobilized yeast population. The settling characteristics of the superflocculent yeast indicate that the yeast population will settle out quite quickly if left stagnant. With a three-phase gas-lift draft tube bioreactor, it will be possible to keep these cells in circulation. However, in the case of a failure in the gas supply system, the cell population would settle out quickly and possibly necessitate auxiliary sparge gas to resuspend the solids. For post-fermentation processing, this rapid settling characteristic is advantageous as solids separation devices such as gravity settlers can be used for bulk solids removal. In the brewing industry, this would decrease the solids load on the centrifuging equipment and therefore allow for longer run times between centrifuge bowl discharges. Less beer loss would be expected, and the level of off-flavors imparted to the beer by the centrifuge (although minimal) would be minimized because of the lower levels of yeast biomass passing through the centrifuge.
Assessment of Gas-lift Technology for Continuous Fermentation of Beer
The first and most important goal of this invention was to evaluate the feasibility of operating the 50-L pilot scale gas-lift bioreactor in continuous mode using κ-carrageenan gel beads to entrap Saccharomyces carlsbergensis cells (described previously). In addition, it was the inventors' desire to investigate whether a North American type lager beer of acceptable flavor quality could be produced with such a system. The inventors also set out to determine the minimum residence time required for complete attenuation of the high gravity wort (17.5° P) as well as establish an operating range for oxygen within the continuous fermentation system.
The minimum residence time where all wort sugars were consumed was 24 hours. This can be compared with a classical batch fermentation time of five to seven days. The dissolved oxygen concentration measured by the in-place Ingold oxygen probe within the bioreactor was close to zero regardless of the oxygen added to the sparging gas (ranged from 0 to 20 percent v/v). This indicated that the oxygen supplied in the wort was either consumed quickly by the yeast cells or was simply vented in the off-gas. The level of free cells in the overflow of beer was in the order of 108 cells per mL of green beer. The levels of the vicinal diketones, diacetyl and 2,3-pentanedione, as well as the level of acetaldehyde, decreased with decreasing oxygen proportions in the sparging gas (
At pseudo-steady state, the pilot scale bioreactor was operated with a residence time of 24 hours over a six-week period. The “green” beer had an acceptable flavor profile and no major defects (sulfury off-notes) were noted. The amount of oxygen in the sparging gas proved to be a critical element in this experimentation. Beers produced with two to five percent oxygen in the sparging gas gave the best taste profiles. This critical control point needs further attention with the focus on more accurate oxygen measurement techniques with measurements performed on a larger set of pre-fermentation and post-fermentation analytes.
In traditional batch primary fermentation, the wort is dosed with oxygen before being transferred into the fermenter. After inoculation of the medium, the dissolved oxygen concentration rapidly decreases as the yeast cells consume it (first 24 hours of fermentation where yeast growth occurs). The remainder of the fermentation is therefore carried out under mostly anaerobic conditions. The use of a continuous homogenous system for primary fermentation does not allow for this change of oxygen concentration over time. For this reason, it may be very difficult to attain a complete flavor match for beer produced using continuous and batch fermentations.
Regardless of these differences, the bioreactor configuration tested in this initial assessment produced a beer with an acceptable flavor quality and analytical profile. By using a gas-lift bioreactor with relatively small-sized beads (˜1 mm), it was possible to increase the volumetric bioreactor productivity by reducing the time for primary fermentation by several days. The level of biomass released in the exiting beer showed that the level of yeast growth in the immobilized cell bioreactor was equivalent to that of free cell batch fermentation under similar conditions. These observations confirm the reliance of flavor formation on the level of yeast growth. This could explain the failure of previous attempts to produce acceptable beer with restricted growth immobilized cell systems. The supply of a controlled gas mixture can be a powerful tool in the fine adjustment of beer organoleptic properties in continuous immobilized cell fermentations.
The high level of diacetyl in the exiting liquid was also observed by other researchers (Virkajarvi & Pohjala, 1999; Kronlof et al., 2000). In North American lager type beers, the target level of diacetyl is 30 μg/L as compared to the 400-800 μg/L levels in the exiting continuous fermentation liquid. The use of traditional aging technology (cold aging at 2° C. for 14 days) lowered diacetyl to the desired range but to the detriment of overall process productivity. The use of the rapid secondary fermentation technology reported in Chapter 2 would help reduce diacetyl without significantly lowering productivity (two-hour process). The additional costs, however, may be prohibitive, and all brewers may not want to subject their beer to high temperatures (80° C. to 90° C.).
Mixing and Circulation Rates of Fermentations within a Three-Phase Gas-Lift Draft Tube Bioreactor
Mixing experiments were conducted using the acid injection method outlined previously. The goals of this phase of the invention were twofold. The inventors wanted to first assess whether the gas-lift system was well mixed by calculating the liquid circulation rate and the resulting mixing time for a number of immobilization carriers at different superficial fluidizing gas velocities. The distinction of these trials from those reported in literature was that the entire experimentation was conducted on actual fermentation media with active yeast cells rather than on model solutions (water-solid-gas systems). These mixing experiments were also geared towards calculating superficial liquid velocities, which could be used for the future scale-up of the pilot system.
Probe Calibration
The pH probe utilized in the mixing studies was calibrated using the method described previously. The 4-20 mA signal emitted from the pH probe was passed across a 250 Ohm resistor and transformed into a 1-5 Volt signal, which could be recorded by the data acquisition card (DAC). The pH probe was immersed sequentially into a series of buffers as follows: pH of 4.6 buffer, pH of 4.0 buffer, pH 5.0 buffer, pH of 6.0 buffer and finally into pH of 7.0 buffer (Beckman certified standards, Cole Parmer).
pH=3.68*Voltage−3.53
The collected mixing data were transformed using the equation to reflect actual pH values as measured inside the gas-lift bioreactor.
The response time of the pH probe to a change in pH was also measured. FIGS. 50 to 53 depict the typical response of the pH probe to four different step changes in pH, namely 0.6, 1.2, 2.3 and 3.4 units. The original data were fitted to a pulse function using the built-in capabilities of TableCurve 2D (Jandel Scientific) data plotting and analysis software. The resulting curve fits had coefficients of correlation greater that 0.994. From these curves, the time for the probe to respond to 98 percent of the step change (response time) was calculated and then plotted versus pH step change (
This type of probe characterization is important, especially when the system is being used for the evaluation of mixing time and circulation rate. The probe should have a low response time for it to reflect changes in the medium it is measuring. In particular, the pH probe response time must be lower than the circulation rate within the reactor if it is to accurately be used in such measurement. A quasi-instantaneous response is however not necessary since a slight lag in response will simply be reflected in consecutive circulation rate measurements and therefore nullified.
Mixing Time and Circulation Rates
Mixing time and circulation rate experiments were performed on three types of immobilized cell fermentations. The experimentation was carried out inside the 50-L pilot scale draft tube bioreactor on a model water solution containing no solids and then on fermentation broth with a specific gravity of 2.7° P containing either κ-carrageenan gel beads, superflocculent LCC290 yeast or medium flocculent LCC3021 yeast.
For the superficial liquid correlation, Kennard and Janekah (1991) proposed an exponent of 0.41 in distilled water and 0.64 when the solution contained carboxymethyl cellulose and ethanol. The LCC290 and the LCC3021 systems had exponents of 0.419 and 0.427 respectively, while the κ-carrageenan system and the water/no solids systems had an exponent of 0.283.
A basic assumption of gas-lift draft tube technology is that the system can deliver adequate mixing so that the fluid element exiting the reactor is completely mixed. In the operation of the 50-L pilot scale systems as continuous fermenters, fresh nutrient medium was injected at the bottom of the reactor at a flowrate of 36 mL per minute into a total reactor volume of 50 L. This represents approximately a 1000-fold dilution in feed components. With the mixing characteristics calculated for the LCC290 yeast system, a fluid element is mixed in about three reactor circulation loops, while 10 reactor circulation loops are necessary for the water/no solids scenario. In addition, the residence time (24 hours) is approximately 1000 greater than the mixing time (180 seconds). The rapid mixing coupled with the dilution of nutrients within the system and the large difference in mixing time versus residence time strongly suggest an adequately mixed system. The original premise for using a gas-lift bioreactor was that it provided an ideally mixed environment for beer fermentation. The results of the mixing studies performed on all three fermentation carriers support this premise.
Evaluation of Several Immobilization Methods for Continuous Primary Fermentation within a Gas-Lift System
Continuous fermentations were performed in the 50-L pilot scale gas-lift draft tube bioreactor utilizing three types of immobilization carrier—κ-carrageenan gel beads, superflocculent LCC290 yeast and medium flocculent LCC3021 yeast. All fermentations were initially pitched with the same level of yeast inoculum (4 g/L) and industrial lager wort was used as the nutrient medium. The fermentations were started up in batch to allow for the rapid reduction of wort sugars, as well as to promote yeast growth within the fermenter. In the case of the flocculent yeast, this batch phase allowed for the formation of yeast flocs, which could then be retained within the bioreactor once a continuous feed was started. Once the diacetyl level in the fermentation liquid had dropped to below 30 μg/L, continuous wort feed was started. The following sections describe in more detail the fermentation analyses resulting from these three types of immobilization matrices.
Use of κ-Carrageenan Gel Beads: Entrapment
The use of κ-carrageenan gel beads as an immobilization matrix aided in the feasibility assessment of gas-lift technology for continuous primary fermentation of beer. It was still necessary to assess whether such a system could be operated for extended periods of time (up to two months) without experiencing major operational difficulties, including fermentation instability and contamination. The operational parameters for this fermentation trial are depicted in
The sudden increase in free yeast concentration is probably linked to the immobilized yeast cell population. At the start of the fermentation, the yeast entrapped in the gel will grow inside the gel until all available space has been taken up. Once the gel beads are filled with yeast, the expanding population will overflow into the liquid medium. The inventors' results seem to indicate that during the first 400 hours, the immobilized yeast grew within the gel and, at ˜700 hours, the yeast had no more room to grow within the bead therefore started releasing larger quantities of cells into the fermentation broth.
This instability in the yeast population is reflected in the ethanol and specific gravity profiles (
Two other key analytes—diacetyl and 2,3-pentanedione—were monitored throughout the two-month continuous run (
Use of a Superflocculent Yeast Strain: Self-Flocculation
Continuous fermentations using the 50-L pilot scale gas-lift bioreactor loaded with LCC290 superflocculent yeast were performed over a three-month period. The CO2 sparging gas was set at ˜2.5 mm/s while air was introduced at ˜0.4 mm/s in order to promote some yeast growth (this corresponds to a total of 1.51 L of gas per minute being introduced with three percent being oxygen). The fermentation temperature in the reactor was maintained at ˜15° C. throughout the entire run. An interruption in the power supply forced the inventors to reduce the temperature of the fermenter to 4° C. for a period of three days (about 1700 hours into the fermentation) (
Once the batch start-up phase was completed (˜180 hours), wort was continuously supplied to the system at a flowrate of 2.16 L per hour, thus providing a 24-hour residence time based on a reactor working volume of 50 L. After the initial batch period, the cell concentration increased, reaching three billion cells/mL at about 750 hours into the fermentation (
The ethanol concentration and the fermentation broth specific gravity reached pseudo-steady state shortly after the continuous feed was started (
The diacetyl and 2,3-pentanedione concentrations over time are presented in
Use of a Medium Flocculent Yeast: Self-Flocculation
Several fermentation runs were performed using the medium flocculent yeast strain LLC302 1, as the immobilization matrix within the 50-L pilot scale gas-lift bioreactor. As with the two previous modes of immobilization, the initial yeast concentration was set at 4 g/L. This yeast was pitched into industrial lager wort (described above) and allowed to ferment in batch until all the fermentable sugars were consumed and the diacetyl level had dropped to below 30 μg/L. The fermentation temperature was controlled at 15° C. and the sparge gas rate was the controlled at the same level as the LCC290 fermentation (superficial carbon dioxide gas velocity of 2.5 mm/s and ˜0.4 mm/s air resulting in approximately three percent oxygen in the total sparge gas) (
This initial batch stage allowed the yeast cells to flocculate and therefore be more easily retained within the gas-lift system. At the end of the batch startup, the wort feed rate was set at 2.16 L per hour, which corresponded to a residence time of ˜24 hours based on a reactor working volume of 50 liters. The yeast population (
Comparison of the Various Carriers
In the previous sections, the fermentation performances of κ-carrageenan gel beads, LCC290 superflocculent yeast and LCC3021 medium flocculent yeast as immobilization matrices were presented. All three carriers were deemed suitable for continuous primary fermentation within the 50-L pilot scale gas-lift draft tube bioreactor. Liquid residence times of 24 hours were achieved in all three cases. The fermentation runs using LCC290 superflocculent yeast reached a pseudo-steady state much quicker than the LCC3021 medium flocculent yeast and κ-carrageenan immobilized cell systems. The LCC290 fermentation reached its maximum ethanol concentration of 70 g/L at about 250 hours into the run. The LCC3021 run hit its steady state ethanol concentration of 70 g/L at ˜600 hours. During the κ-carrageenan continuous fermentation, ethanol leveled off at two separate points over the course of the run. First, ethanol hit 45 g/L between 200 and 500 hours and then rose to 70 g/L at about 575 hours and remained at that concentration until the end of the trial.
The three fermentation systems seemed to reach a maximum free yeast cell concentration of ˜1 billion cells per milliliter. The inconsistency in yeast concentration impacted negatively on the κ-carrageenan system's ethanol productivity (lower initial pseudo-steady state ethanol concentration as compared to LCC290 yeast system). The yeast concentration peaked at different time intervals in each system. For the LCC290 run, ethanol reached its maximum between 500 and 1000 hours into the continuous fermentation, while the LCC3021 fermentation had a maximum cell count between 1500 and 2000 hours into its continuous run. The κ-carrageenan immobilized system reached a maximum cell concentration between 700 and 1000 hour of continuous operation.
The pseudo-steady state concentrations of diacetyl and 2,3-pentanedione in the three types of immobilized cell fermentation—LCC290 superflocculent yeast, LCC3021 medium flocculent yeast and κ-carrageenan immobilized yeast—were dissimilar. For the LCC290 fermentation, diacetyl settled at 375 μg/L while the level in the LCC3021 fermentation was leveled off at about 225 μg/L. In the case of the κ-carrageenan fermentation, the diacetyl concentration hit 500 μg/L, and midway into the continuous run, the level decreased gradually to about 200 μg/L in a 500 hour time frame. The 2,3-pentanedione concentration mirrored the diacetyl concentration in all three runs with concentrations of 2,3-pentanedione higher than diacetyl throughout the LCC290 and LCC3021 fermentations. The κ-carrageenan run exhibited a different pattern, with diacetyl levels higher than 2,3-pentanedione during its first pseudo-steady state, after which time the diacetyl concentration dropped below the 2,3-pentanedione concentration. The yeast concentration data and the ethanol production data also suggest that two separate and unique pseudo-steady states were achieved during the κ-carrageenan fermentations.
The task of comparing different fermentation systems and assessing which one has performed better can become complex when the merits of the system are based on more than one criterion. For example, if overall ethanol productivity alone was used as the measure of success, all three tested systems would rate equally well since the production of 70 g/L ethanol per 50 L of reactor volume over a 24-hour residence time was achieved in all cases.
The production of a salable beer requires more than simple ethanol production. The proposed fermentation system should be evaluated on its ability to produce an acceptable beer (ethanol productivity and diacetyl levels among other things), on the potential incremental costs of the carrier, on the availability of the carrier, on the ease of operation of the system, on environmental issues such as disposal of the carrier, on the stability of the carrier, as well as the flexibility provided by the carrier system. In order to evaluate such a multifaceted scenario, the business world utilizes a dimensionless analysis process called the “Balanced Scorecard” (Kaplan and Norton, 1996). The first step involves the identification of criteria on which the system must be evaluated. Each criterion is then given a rating on a scale of one to five, with one being least favored and five being the most favored. At the end of the analysis, the score for each option is totaled and the alternative with the highest score is the best choice given the circumstances.
Table 6 presents the results of the Balanced Scorecard analysis performed on the immobilization carriers that were viewed as potential alternatives for use in the 50-L pilot scale gas-lift draft tube bioreactor for fermentation. A total of six carriers—CHITOPEARL® chitosan beads, CELITE® diatomaceous earth beads, SIRAN® glass beads, κ-carrageenan gel beads, medium flocculent LCC3021 yeast and superflocculent LCC290 yeast—were evaluated with the primary objective of producing a salable beer. Each carrier system was rated using the aforementioned scale. Overall, the LCC290 superflocculent yeast performed best followed closely by the medium flocculent yeast, LCC3021. The four other carriers received scores between 16 and 20. Third preference was given to the κ-carrageenan system because a salable liquid was produced in the pilot unit.
This carrier assessment strongly suggests that future focus on the development of continuous fermentation systems should be geared toward self-aggregation as the mode of immobilization. The availability (readily available), cost (low cost since no additional equipment is needed to operate), ease of operation (fits within existing plant operations) provided by such an alternative outweigh the potential instability of the yeast flocs in an agitated system. It may be possible to use the shear sensitivity of self-aggregate to control the floc size during the fermentation process, and possible to achieve increase mass transfer and therefore achieve further increases in bioreactor volumetric productivity.
6.6 Production of a North American Type Lager Beer Using Gas-Lift Draft Tube Technology
The production of a clean tasting North American (NA) type lager beer poses many challenges to the brewer. NA type lager beers are characterized by a light color and a taste profile with low bitterness, low residual sugar (thin), no dominant flavor and therefore relatively no aftertaste. Because of these inherent properties, the brewer can mask very few flavor defects. High levels of diacetyl (buttery), acetaldehyde (green apple) as well as sulfury off-notes (burnt rubber, skunky, rotten eggs, cooked vegetables) are the most common flavor problems plaguing modern day brewers. Although bacterial contamination of the fermentation medium can also be a cause of these off-flavors in beer, improper control of the fermentation process more often yields higher than expected off-flavor levels.
Throughout the continuous fermentation trials conducted as part of this invention, contamination levels in both the wort and the fermentation vessels were controlled through diligent practice of aseptic techniques. The fermentation trials with all three carrier types, which lasted for several months, showed no detectable levels of contaminants. Higher than desired levels of diacetyl (target less than 30 μg/L) and acetaldehyde (target less than 10 mg/L) plagued the products from the continuous primary fermentations, but these levels were not due to bacterial infection. These findings do not differ from those reported in literature (Pajunen et al., 2000; Kronlof et al., 2000). A Belgian brewer turned its high levels of acetaldehyde in the beer from a continuous fermentation process into a selling feature and marketed the product as an apple-flavored ale (Andries et al., 1996b).
High levels of diacetyl following primary fermentation are also normal in the brewing industry. Some brewers have adopted a practice called “temperature free rise” following the completion of their primary fermentation in order to aid in the reduction of diacetyl. Others have opted to simply hold their products for longer periods of time during the aging process to achieve the reduction of the vicinal diketones (diacetyl and 2,3-pentanedione) to the desired levels. In another approach, several research groups have developed rapid aging technology to deal with high diacetyl levels. Although this approach is very effective, it adds another level of complexity to the overall brewing process that some may find difficult to accept.
The economics of the rapid aging process are quite clear. However, in these early days of continuous processing in the brewing industry, it may be advisable to minimize technological complexity in order to facilitate the transition from traditional batch fermentation to continuous production. For this reason, it was decided to investigate the use of a batch hold following continuous primary fermentation within the 50-L pilot scale gas-lift systems to control the high levels of diacetyl in the finished beer. This additional processing step was not foreseen at the beginning of this invention. However, it was necessary to implement such a measure in order to compare the beers produced continuously to batch control beers.
Use of Batch Hold Following Continuous Primary Fermentation
A critical parameter in determining the completion of primary fermentation is the diacetyl level in the end fermented liquid. The conversion of the diacetyl precursor, α-acetolactate, into diacetyl is the rate-limiting step in the diacetyl pathway. This first reaction is chemical in nature and is highly dependent on temperature. If the “green” beer enters the cold aging process before the chemical conversion of α-acetolactate to diacetyl has occurred, the resulting beer may have levels of diacetyl above the taste threshold of 20 μg/L, unless extended cold aging periods are used to allow the slow conversion of the precursor to occur. In all three continuous fermentations described in section 6.5, the diacetyl level exiting the reactor was above the desired target value of 30 μg/L in the undiluted beer. If the liquid was filtered at this stage to remove the yeast, the diacetyl would remain high. Hence a warm batch period was employed to reduce the diacetyl value to below the acceptable limit.
The continuously fermented beer was collected and held in 40-L stainless steel beer vessels at 21° C. Small samples (100 mL) were withdrawn regularly from the liquid and analyzed for diacetyl, ethanol, specific gravity, esters and fusel alcohols.
In another test, the effect of agitation on the reduction of diacetyl during the hold period was investigated. A carbon dioxide sparge gas (0.14 m3/h) was introduced through a 1.27 cm stainless steel tubing into the bottom of the beer collection vessel to keep the liquid agitated during the hold period.
The effects of this warm batch hold following continuous primary fermentation on the concentrations of esters and fusel alcohols and on the ethanol concentration and the specific gravity of the liquid are presented in
The batch holding protocol was performed on liquid produced from the continuous fermentations in the 50-L gas-lift bioreactor using LCC290 superflocculent yeast, LCC3021 medium flocculent yeast or κ-carrageenan immobilized yeast. The diacetyl reduction profiles of these three test runs are presented in
[Diacetyl]=648.54 e(−0.0426 t)
with a correlation of coefficient of 0.96. This result strongly supports the theory that all three immobilization systems exhibited the same diacetyl reduction potential. The results should not be surprising since diacetyl reduction is often linked to yeast strain (Nakatani et al., 1984). The κ-carrageenan system immobilized the LCC3021 yeast within its gel structure and the LCC290 yeast was a selected variant of the LCC3021 strain.
Once the diacetyl level was below the target level of 30 μg/L, the resulting beers were aged in cold storage (2° C.) for seven days before undergoing final product preparation (filtration, dilution, carbonation and packaging). Table 7 summarizes the analysis of the beers produced in the 50-L pilot scale systems with either LCC290 yeast, LCC3021 yeast or k-carrageenan immobilized yeast as the immobilization matrices.
Several other parameters (apparent extract, real extract, calculated original extract, color, bitterness) that are intimately affected by the dilution of the product from its original ethanol concentration to the final value of 5.0 percent v/v were different from the control because the continuous products required higher dilution with water to reach the desired ethanol value due to higher original ethanol concentrations (70 g/L versus 60 g/L in batch).
Selection of Best Sparge Gas
Carbon dioxide is readily available in most breweries since it is a natural by-product of yeast fermentation. Breweries collect the evolved CO2 and then scrub the gas stream to remove slight impurities that may have been carried over into the collection stream (typically sulfurous compounds). This purified stream is then compressed and stored as a liquid for future use in the brewery (99.95 percent pure). The use of carbon dioxide as a sparge gas in continuous fermentation seemed like a logical choice from an operations point of view. The plant would be able to utilize their collection system and recover the CO2 exiting the continuous fermenter. The use of other gases would only add another level of complexity to the existing plant operations.
However, for continuous fermentation to become a viable alternative to existing batch operations, it is necessary to produce a product that is a close match to existing brands. It is believed that by minimizing the biological/biochemical impacts that the yeast is exposed to during continuous fermentation, it may be possible to achieve such a product match. Carbon dioxide is known to adversely affect yeast metabolism during primary fermentation. This effect is magnified in tall cylindroconical vessels where the head pressure inherent to the system suppresses the free release of CO2 from the liquid medium. These conditions tend to produce beers with lower esters and higher fusel alcohols. Successful attempts have been made to remove some of this CO2 from the fermentation by periodically injecting an auxiliary gas stream into the bottom of the cylindroconical vessel. The effects of CO2 inhibition appeared to have been reduced with the resulting products containing less fusel alcohols and higher ester levels.
Building on this knowledge, the use of nitrogen gas instead of carbon dioxide, as the sparge gas within the gas-lift system, was investigated. Several liquids fermented in the 50-L uplift reactors were collected and processed in the Labatt Experimental Brewery utilizing the following steps. After 14 hours of liquid collection from the reactor (24 hour residence time), the “green beer” was decanted from the yeast. This liquid was then subjected to a 48 hour room temperature (21° C.) hold period, which allowed both diacetyl and acetaldehyde to reach Labatt specifications (diacetyl<30 μg/L and acetaldehyde<10 mg/L). This liquid was then cold aged for seven days and then processed using standard industrial practices.
Table 8 compares the results of finished beers obtained from continuous fermentation with LCC290 yeast under CO2 sparged and N2 sparged systems to a standard industrially produced liquid (control). The nitrogen-sparged liquid compares favorably to the industrially produced liquid. Analyses indicated that there was twice as much 1-propanol in the liquid while the dimethyl sulfide concentration was approximately three times lower. Both color and bitterness values appeared to be higher than the industrial liquid as did the foaming potential as measured by the NIBEM test. The CO2-sparged beer had lower esters (ethyl acetate, isoamyl acetate) and higher 1-propanol than did the nitrogen-sparged liquid. The ratios of esters (ethyl acetate, isoamyl acetate, ethyl hexanoate, ethyl octanoate, ethyl decanoate) to fusel alcohol (1-propanol, isobutanol, isoamyl alcohol) were calculated for the control liquid, the nitrogen-sparged liquid and the CO2-sparged liquid and were found to be 0.30, 0.27, 0.15 respectively.
The findings suggested that nitrogen sparging increased the production of esters to similar levels as those of commercial beers, whereas carbon dioxide sparging produced liquids with relatively lower ester concentrations. These results suggest an altered yeast metabolism under the CO2-sparged environment. The propanol levels, regardless of the sparging gas, were much higher in the continuously fermented beers than those measured in industrially fermented control beers. Although the propanol concentrations are well below the taste threshold of 100 mg/L, the noted differences may be an indicator of slightly altered metabolism occurring in continuous fermentation as compared to batch fermentation. It is also possible that the higher propanol level is due to the continuous supply of the amino acid threonine, which through the oxo-acid degradation pathway will yield propanol.
An acceptable beer with no major flavor defects can be produced using a 50-L pilot scale gas-lift draft tube bioreactor as the continuous primary fermentor when followed by a two-day batch hold for the control of diacetyl. A minimum residence time of 24 hours or one reactor volume turnover per day is achievable for the fermentation of high gravity wort (17.5° P) into an end-fermented broth (2.5° P). The use of superflocculent yeast LCC290, medium flocculent yeast LCC3021 and κ-carrageenan immobilized yeast are all feasible carriers within the gas-lift system. The use of heavier pre-formed carriers such as SIRAN® glass beads and CELITE® diatomaceous earth beads are not practical alternatives within a gas-lift draft tube system. A minimum of two months' continuous operation in the case of κ-carrageenan gel beads and a minimum of three months' continuous operation in the case of LCC290 and LCC3021 fermentations are achievable without experiencing any microbial contaminations or reactor performance instabilities. In addition, the continuous system was capable of handling potential and assumed variations in the industrial wort supply during the extended run periods.
Continuous fermentation using the superflocculent yeast LCC290 with nitrogen as the sparging gas, followed by a two-day batch hold, produced the closest flavor match to an industrial control beer. The two-day batch hold devised to deal with the high concentrations of diacetyl in the outlet liquid of the continuous primary fermentor was an effective, although not optimal, control mechanism. The diacetyl reduction capacity of the three tested continuous fermentation systems was very similar and, as previously suspected, this trait can be attributed to strain type. The batch hold period did not affect the concentrations of esters and fusel alcohols in the beer during the holding stage.
Utilizing three percent oxygen in the sparge gas provided adequate oxygen nutrient levels in the wort, resulting in the maintenance of a viable yeast population (above 90 percent) throughout the fermentation runs while producing beers with an acceptable flavor profile. The continuous fermentations using LCC290 yeast and LCC3021 yeast as the immobilization matrices reached a pseudo-steady state more rapidly than the κ-carrageenan gel bead system. The instability of κ-carrageenan immobilized fermentations, possibly resulting from an increase in the immobilized yeast population, caused the fermentation of product below target levels. For ideal continuous production, this phenomenon is highly undesirable as the prolonged startup increases the time necessary to react and recommence following a catastrophic failure. A longer startup phase will also require a longer continuous run phase in order to become attractive.
The continuous gel bead production process produced the required quantities of beads for testing inside the pilot scale units. It is, however, necessary to further optimize the bead production process so as to produce beads with a tighter size distribution. The process also requires further investigation to determine its suitability at an industrial scale. Rather than the Kenics type investigated in this research, a new type of static mixer for which scale-up can be achieved by increasing diameter rather than solely on number of static mixers, must be found and tested for this option to become viable.
The acid pulse tracer technique utilized in this invention allowed the inventors to assess the mixing time and circulation rate within the 50-L pilot scale bioreactor during actual fermentations involving LCC290 superflocculent yeast, LCC3021 medium flocculent yeast and κ-carrageenan immobilized yeast. The mixing data were fit to a decaying sinusoidal function from which the mixing time and circulation rate were calculated.
Rapid mixing is provided within the gas-lift draft tube system with mixing times calculated at under 200 seconds for all three types of immobilization carriers. Mixing time decreased slightly with increases in superficial gas velocity in all three tested scenarios. At the all tested superficial gas velocities (2 mm/s to 6 mm/s), the LCC290 system showed the quickest mixing times (between 100 s and 120 s). The liquid circulation times were very similar for all three carrier types regardless of superficial gas velocity. They also decreased linearly with corresponding increases in gas velocity. Complete liquid mixing (98 percent response to a pulse) occurred within three to six reactor circulation cycles for all the tested immobilization carriers. These results confirmed that the tested 50-L pilot scale system provided adequate mixing for continuous fermentation. Poor mixing times would have been indicative of possible dead zones that would have been undesirable for beer fermentation.
This invention clearly demonstrates the feasibility of pursuing a further scale-up of the production system designed, built, and operated in the Labatt Experimental Brewery. The use of a gas-lift draft tube bioreactor with LCC290 superflocculent yeast and nitrogen as the sparging gas, followed by a two-day batch hold at 21° C., is the system of choice.
Detailed Description—Part 2:
Materials and Methods
Yeast Strain and Characteristics
A lager brewing strain of Saccharomyces cerevisiae, Labatt Culture Collection (LCC) 3021, was used throughout this invention. Saccharomyces cerevisiae is synonymous with Saccharomyces uvarum Beijerinck var. carlsbergensis Kudryavtsev, 1960 (Kurtzman, 1998). At 37° C. LCC 3021 will not grow. This helps to distinguish LCC 3021 lager yeast from most ale yeast, which will grow at 37° C. and higher temperatures. LCC 3021 is a bottom fermenting strain, as are most lager yeasts, but there are exceptions. As well, this strain will ferment glucose, galactose, sucrose, maltose, raffinose, and melibiose, but not starches. The ability to ferment melibiose is one tool used by taxonomists to distinguish it from ale yeast.
As with most brewing strains, LCC 3021 is polyploid and reproduces by mitotic division. Under normal brewing conditions lager yeast does not reproduce by meiosis. This has the advantage of making the brewing strain genetically stable because crossover of genetic material is less likely (Kreger-van Rij, 1984).
Preparation of Yeast Inoculum
Yeast was taken from a vial cryogenically preserved in a −80° C. freezer and streaked on Peptone Yeast-Extract Nutrient (PYN) agar (peptone, 3.5 g/L; yeast extract, 3.0 g/L; KH2PO4, 2.0 g/L; MgSO4.7H2O, 1.0 g/L; (NH4)2SO4, 1.0 g/L; glucose, 20.0 g/L; agar, 20.0 g/L in dH2O) growth medium to obtain well-separated colonies. A sterile loop consisting of several colonies was taken from the three-to-four day old plate of growing yeast, and these colonies were inoculated into a 10 mL volume of wort in a test tube. This was allowed to grow at 21° C. overnight, thus the term “overnight culture.” It was then added to a larger volume of wort, usually 200 mL, to increase yeast biomass. In consecutive days, this mixture was added to another larger volume of wort, and so on, until the desired amount of yeast biomass was propagated. Generally one expects to produce approximately 20 grams of lager yeast per litre of wort. To prepare for yeast inoculation, the culture was centrifuged at 4° C. and 1.0×104 rpm (radius=0.06 m) for 10 min. After centrifuging, the liquid was decanted and the appropriate wet weight of yeast was obtained from the pellet for pitching.
Wort Fermentation Medium
A brewery wort with a specific gravity of 17.5° P was used in this invention. Additional detail on the wort composition is given by Dale et al., 1986, Hoekstra, 1975, Hough et al., 1982, Klopper, 1974, and Taylor, 1989.
Batch Fermentations: Wort was heated in an autoclave for 45 minutes at 100° C. and then cooled, before inoculation with immobilized cell beads or freely suspended yeast.
Continuous Fermentations: The wort used for the continuous fermentations was flash pasteurized (Fisher Plate Heat Exchanger, combi-flow Type Eurocal 5FH) prior to feeding into the gas lift bioreactor. This wort was monitored regularly for microbial contaminants, as described in the following sections. If contamination was detected in the wort, it was immediately discarded and new wort was collected from the plant.
The flash pasteurizer was operated at a volumetric flow rate of 0.8 m3/hr. The unit had a tubular holding section where the wort was held at an average temperature of 85° C. with a minimum temperature of 80° C. The volume of the holding section was 1.13×10−2 m3, giving a residence time in the holding section of 51 seconds. Following the heating step, the wort was rapidly cooled to a temperature of 2° C. upon exiting the unit.
Immobilization Methodology
Kappa-carrageenan gel beads containing entrapped lager yeast cells were produced using the static mixer process, as described in detail by Neufeld et al. (1996), with initial cell loadings of 107−108 cells/mL of gel, which are specified for each experiment. As illustrated in
Cumulative Particle Size Distribution of Kappa-Carrageenan Gel Beads
Containing Immobilized Yeast Cells
Kappa-carrageenan gel beads were randomly sampled from a 30-L production run of gel beads in order to calculate a particle size distribution on a mass wet-weight basis. Each sample was approximately 500 grams wet weight. Sieving was used to determine the bead particle size distribution. The beads were passed through a series of sieves with grid sizes of 2.0, 1.7, 1.4, 1.18, 1.0, and 0.5 mm. A 4.5 L volume of 22 g/L KCl solution was used facilitate the sieving of each bead sample. The kappa-carrageenan gel beads were assumed to be perfectly spherical so that the sieve diameter was taken as the particle diameter. It was also assumed that the particle density was uniform and independent of particle size.
Yeast Cell Enumeration and Viability
Freely Suspended Yeast Viability and Cell Concentration: The American Society of Brewing Chemists International methylene blue staining technique (Technical Committee and Editorial Committee of the ASBC, 1992) was used to measure yeast cell viability. The stain measures whether a yeast population is viable or non-viable based on the ability of viable cells to oxidize the dye to its colorless form. Non-viable cells lack the ability to oxidize the stain and therefore stain blue. Fink-Kuhles buffered methylene blue was prepared by mixing 500 mL of Solution A (0.1 g methylene blue/500 mL dH2O) with 500 mL of Solution B (498.65 mL of 13.6 g KH2PO4/500 mL d H2O mixed with 1.25 mL of 2.5 g Na2HPO4.12H2O/100 mL d H2O) to give a final buffered methylene blue solution with a pH of 4.6.
The diluted yeast solution was mixed with the methylene blue solution in a test tube, to a suspension of approximately 100 yeast cells in a microscopic field. A small drop the well-mixed suspension was placed on a microscope slide and covered with a cover slip. Following one to five minutes of contact with the stain, the cells stained blue and the cells remaining colorless were enumerated. The percentage of viable cells was reported as a percentage of the total number of cells enumerated. Cell concentration was determined using a light microscope and a Hemacytometer (Hauser Scientific Company).
Immobilized Cell Viability and Cell Concentration: Gel beads were separated from the fermenting liquid by passing the mixture through a sterile sieve (500 μm pore mesh size) and rinsing with 10 mL of distilled water. Gel beads, 1 mL, containing entrapped yeast were added to a sterile 50 mL specimen container containing 19 mL of distilled water. The beads were then disrupted using a POLYTRON® (Brinkmann Instruments) apparatus, to release the cells from the gel. Cell viability and concentration were then measured as described for the freely suspended cells.
Microbiological Analyses
Liquid Phase Analyses: Samples were taken from continuous fermentations at least once a week for microbiological analyses. The wort that was used for continuous fermentations was also tested for contamination prior to transferring it into the bioreactor. To test for the presence of both aerobic and anaerobic bacteria, samples were plated on Universal Beer Agar (UBA, Difco Laboratories), with the addition of 10 mg/L of cycloheximide, and incubated at 28° C. for 10 days. Plates that were tested for anaerobic bacterial contamination were placed in an anaerobic jar with an ANAEROGEN® (Oxoid) packet, which takes up any oxygen remaining in the jar, creating an anaerobic environment. An anaerobic indicator (Oxoid), which turns pink in the presence of oxygen, was used to verify anaerobic conditions within the jar. Wild yeast contamination was tested by plating samples on yeast medium (YM agar, Difco Laboratories) plus CUSO4 (0.4 g/L) incubated at 25° C. for seven days. Peptone Yeast-Extract Nutrient agar (PYN) , described previously, was used to screen samples for non-lager yeast contaminants at 37° C. for seven days. The absence of yeast growth on PYN at 37° C. indicated that no ale yeast or contaminants that grow at 37° C. were present.
Gel Phase Analyses: An assay was developed in the inventors' laboratory to ensure that the immobilized cell beads to be used for fermentations were free of contaminating bacteria before being pitched into the bioreactor. The main concern was to avoid contamination with beer spoilage organisms such as Pediococcus sp. and Lactobacillus sp. or wild yeast. A 3 mL volume of carrageenan gel beads was inoculated into 100 mL of several different selective liquid media described below, placed in 250 mL flasks at 25° C., and shaken at 100 rpm in an incubator shaker. NBB broth (Nachweis von Bierschädlichen Bacterien) (BBL cat # 98139, NBB Broth Base, 0.02 g/L cycloheximide) is a semi-selective medium which is used to test for beer spoilage bacteria, such as Pediococcus sp. and Lactobacillus sp. Copper sulphate broth (16 g/L YM broth, Difco; 0.4 g/L CuSO4) is a semi-selective medium to test for wild yeast contaminants. Finally, Standard Methods (STA)+cycloheximide broth (16 g/L “Standard Methods” broth, Difco; 0.02 g/L cycloheximide) is used to test for bacteria found in water, wastewater, dairy products, and foods (Power and McCuen, 1988). The selective media were chosen to detect and identify potential beer spoilage organisms within three days. Contaminated samples were indicated by turbidity within the sample and a presumptive identification of the contaminants was made.
Respiratory Deficient (RD) Yeast Cell Detection Methodology: Triphenyltetrazolium Choride (TTC) Overlay Technique: This method was used to distinguish respiratory deficient yeast from the rest of the population and is based on the principle that TTC is a colorless salt that forms a red precipitate upon reduction. When TTC is overlayed onto yeast colonies growing on Yeast-Peptone-Dextrose (YPD) agar (yeast extract, 10 g/L; Peptone, 20 g/L; Dextrose, 20 g/L; Agar, 20 g/L in dH2O), respiratory sufficient yeast will reduce the TTC, and these colonies will become dark pink to red. However, respiratory deficient yeast do not reduce the dye and retain their original color.
Cultures were serially diluted to a suitable concentration of microorganisms, ˜100 cells/0.2 mL, for plating. The YPD plates were then incubated for approximately three days at 21° C. until yeast colonies were visible in an aerobic environment. Each plate was then overlaid with 20 mL of 50° C. TTC overlay agar. After cooling the individual solutions to 50° C., TTC overlay agar was made by mixing 1: I Solution A (12.6 g/L NaH2PO4; 11.6 g/L Na2HPO4; 30.0 g/L agar in dH2O, autoclaved at 121° C., 15 min) with Solution B (2.0 g/L 2,3,5-triphenyltrazolium chloride in dH2O, autoclaved at 121° C., 15 minutes). Plates were read after three hours of incubation at ambient temperature. Percent RD was reported as a percent of unstained colonies of the total number observed.
Scanning Electron Microscopy (SEM) of Yeast Immobilized in Kappa-Carrageenan Gel Beads
Kappa-carrageenan gel beads containing immobilized yeast were removed from the bioreactor through the sample port and placed in a 10 mL screw-cap glass vial, with the beads submerged in a small volume of fermentation broth. The vial was immediately covered in ice and transported in an insulated container to the SEM facility. Kappa-carrageenan gel beads containing immobilized yeast were fixed in two percent (v/v) glutaraldehyde prepared in Sorensen's phosphate buffer, 0.07 M, pH 6.8 (Hayat, 1972). This was followed by post-fixing in one percent (w/v) osmium tetroxide, prepared in the same buffer, and dehydration through a graded series of alcohol solutions 50, 70, 80, 90, 95, 100 percent (v/v), at 15 minutes for each, and then three changes at 100 percent. Before critical point drying (Ladd Research Industries, Burlington, Vt.) through carbon dioxide, some beads were frozen in liquid nitrogen, fractured, and collected into 100 percent alcohol. Freeze fracturing allows the internal face of the beads to be exposed with minimum distortion. Following critical point drying, the samples were sputter-coated (Polaron SC500 sputter coater, Fison Instruments, England ) with 30 nm of gold/palladium and then scanned with a Hitachi S-4500 field emission scanning electron microscope (Nissei Sangyo, Tokyo, Japan).
Bioreactor Sampling Protocol
The bioreactor sample port (SCANDI-BREW® Type T Membrane Sample Valve) reservoir was filled with 70 percent (v/v) ethanol solution to maintain aseptic conditions around the opening between samplings. In order to take a sample, the plug was removed from the base of the ethanol reservoir, drained, and rinsed thoroughly with ethanol, prior to opening the port. Samples were collected into a crimp vial or a screw-cap jar and volumes varied from 5-60 mL, depending on the analysis required. In order to test for microbiological contamination, 10 mL of the fermentation liquid was vacuum-pumped though a sterile membrane filter unit. The membrane, 0.45 μm pore size, was placed on the appropriate selective medium, as described previously.
For chemical analyses, 60 mL of sample was withdrawn through the septum from the 100 mL crimp-sealed vial and syringe-filtered through a Schleicher and Schull, FP-050, double-layer syringe filter system, 5 μm and 0.45 μm pore sizes. The required volume of sample was then dispensed into a 20 mL head space vial and crimped with a Teflon septum and aluminum cap. The required sample volumes are listed in Table 9.
Dissolved Oxygen Measurement
The electrochemical measurement method used by the dissolved oxygen analyzer is based on an amperometric three-electrode arrangement with a potentiometer. The measuring cell consists of a measuring electrode (cathode) and counter electrode (anode). These electrodes are exposed to the liquid in which the oxygen concentration is to be measured. A reaction at the measurement electrode occurs after fixing a defined measurement potential. At the large, silver, measurement electrode, molecules of oxygen are reduced to hydroxyl ions. Two water molecules react in the equation below, with one molecule of oxygen, while absorbing four electrons, giving four hydroxyl ions.
O2+2H2O+4e−→4OH−
The stainless steel anode absorbs the four electrons released at the cathode in order to ensure the flow of current. In the next equation, the measurement current, I, is directly proportional to the oxygen concentration, CL,O:
I=K×CL,O
where the constant, K, in influenced by the Faraday constant, the number of electrons converted per molecule, the cathode surface area, and the width of the boundary layer at the surface of the measurement electrode.
A constant, characteristic, measurement potential is critical for the selectivity (for oxygen) and precision of the measurement. The measuring voltage is stabilized by the reference electrode, which is not burdened by current. This, together with the potentiostat, which provides electronic feedback, provides a constant measurement potential. The surface of the measurement electrode is electrolytically connected to the reference electrode via a diaphragm.
The error, based on the measuring range of the final dissolved oxygen concentration, was ±3 percent (Anon, 1998). The dissolved oxygen analyzer was calibrated using the Thiedig Active Calibration, in which the dissolved oxygen analyzer produced a defined oxygen quantity based on Faraday's Law (0.500 mg/L) and then cross-checked this with the measured values in the matrix. This allowed the instrument to be calibrated under the pressure, temperature and flow conditions corresponding to those of the measurement, within one minute. Because the exchange of molecules in the sensor is a diffusion process, it is influenced by temperature, resulting in faster reaction rates and increases in the measured current. Therefore, the dissolved oxygen analyzer is also equipped with a sensor, which measures the temperature and automatically compensates for fluctuations.
Dissolved Oxygen Measurement in the Wort During Storage: Food grade tubing (¼ inch i.d.) was aseptically connected to a sample port located near the top of the conical bases of the wort storage tanks, T-1 and T-2 (see section 4.2.1). A variable speed peristaltic pump provided volumetric flow rate of 11 L/hr through the dissolved oxygen analyzer block. Wort dissolved oxygen measurements were then recorded after four to five minutes.
Dissolved Oxygen Measurement in the Bioreactor: Prior to performing the dissolved oxygen measurements on the bioreactor, the dissolved oxygen analyzer block was sanitized. The inlet of the sensor was connected to sterile, food grade tubing (¼ inch i.d.). A 70 percent (v/v) ethanol solution was pumped through the analyzer at a volumetric flow rate of approximately 10 L/hr for 15 minutes. The dissolved oxygen analyzer was connected to a laboratory water tap, and hot water (70° C.) was passed through the sensor for a minimum of two hours. This methodology was used rather than steam sterilization because the analyzer block materials cannot tolerate temperatures of above 70° C. Following the two-hour sanitation period, the tubing at the inlet and outlet of the unit was clamped to maintain sterility within the analyzer. In a laminar flow hood, the freshly sterilized tubing was connected to the inlet and outlet of the analyzer. The free ends of the tubing were then aseptically clamped to the ¼″ I.D. stainless steel ports on the bioreactor head plate, and measurements were taken. When the ports on the bioreactor were not in use, they were sealed using a short length of sterilized food grade tubing.
Dissolved oxygen was measured on-line in the gas lift bioreactor by withdrawing liquid from the fermentation through a port situated on the bioreactor head plate. The fermentation liquid exited the bioreactor through a stainless steel filter (see section 4.1.2) connected to a ¼ inch stainless steel pipe which penetrated the bioreactor head plate. The liquid then flowed through flexible food grade tubing (¼ inch i.d.), which was connected to a variable speed peristaltic pump, providing a volumetric flow rate of 11 L/hr through the dissolved oxygen analyzer block. The fermentation liquid was then recycled through a second quarter-inch stainless steel port, which penetrated the bioreactor head plate. Food grade tubing (Cole-Parmer, 1999) was used to connect the sensor to the bioreactor because of its supplier-specified low oxygen permeability of 30 cm3mm/(s·cm2·cmHg)×10−. The measurement was taken after four to five minutes of circulation.
Chemical Analyses
Calibrations were performed using the appropriate standard reagents. All reagents used for the analyses were >99 percent pure. Where necessary, subsequent purification via distillation was performed.
Ethanol
Ethanol concentration was determined using the internal standard gas chromatagraph (GC) method of the Technical Committee and Editorial Committee of the American Society of Brewing Chemists (1992). Degassed samples were treated directly with isopropanol internal standard, five percent (v/v) and injected into a gas chromatograph equipped with a flame ionization detector (FID) and a Dynatech autosampler. A Chromosorb 102, 80-100 mesh column was used with helium as the carrier gas. Chromatographic conditions: flow rate of 20 mL/min, injector temperature of 175° C., detector temperature of 250° C., and column temperature of 185° C.
Carbohydrate Summary
Glucose, fructose, maltose, DP3 (maltotriose), DP4 (maltotetraose), poly-1 (polysaccharide peak 1) and glycerol concentrations in fermentation samples were quantified using a high performance liquid chromatograph (HPLC) equipped with a cation exchange column and a refractive index detector. The mobile phase was potassium phosphate, dibasic, 0.01 M, and the system was equipped with an auto sampler. The instrument was operated with a backpressure of 800 psi. The flow rate of sample and eluent through the column was 0.6 mL/min, with a column temperature of 85° C. and a detector temperature of 40° C. The injection volume was 10 μL.
Specific Gravity
The specific gravity of the wort and fermentation medium is described in this study in terms of Real Extract (degree Plato, ° P), which is the accepted unit used in the brewing industry.
Fermentation samples were filtered as described previously and vortexed prior to analysis with a digitalized density meter to measure wort specific gravity (degree Plato). The fermentation samples were inserted into a glass u-tube, which oscillated electronically to determine the specific gravity, thus giving degree Plato indirectly.
Degree Plato refers to the numerical value of a percentage (w/v) sucrose solution in water at 20° C. whose specific gravity is the same as the wort in question. Because the degree Plato scale and resulting tables relating solution specific gravity to solute concentrations are based on aqueous solution of sucrose, it is only an approximation of the amount of extract. Extract is a term referring to the total available soluble mass in a brewing material “as is,” and /or potentially through processing (Hardwick, 1995) such as carbohydrates, proteins, tannins. Extract is still currently expressed in the brewing industry as specific gravity in degree Plato because of the lack of a more appropriate reference better related to the variability in compositions of worts of different origins.
Total Diacetyl
Total diacetyl (2,3-butanedione) in beer and fermentation samples was measured using a headspace analyte sampling technique, followed by capillary GC separation and electron capture detection (ECD) based on the method of the Technical Committee and Editorial Committee of the American Society of Brewing Chemists (1992). The method refers to “total diacetyl” because the method measures the amount of diacetyl and its precursor, alpha-acetolactate. The carrier gas was five percent methane in argon at 1.0 mL/min, and a wax column was used. The split ratio was 2:1, and the auxiliary gas was helium at 60 mL/min. Injector temperature was 105° C., and detector temperature was 120° C.
The system was equipped with a headspace autosampler and 2,3-hexanedione was used as an internal standard. The sample cycle time was 40 minutes, with a vial equilibration time of 30 minutes at 65° C., a pressurization time of two minutes at 4.8 psig, a loop fill time of 0.2 minute, a loop equilibration time of 0.1 minute, and an injection time of 0.27 minute. Carrier pressure was 18.8 psig, transfer line temperature was 95° C. and loop temp was 65° C.
Beer Volatiles
Beer volatiles including acetaldehyde, ethyl acetate, isobutanol, 1-propanol, isoamyl acetate, isoamyl alcohol, ethyl hexanoate, and ethyl octanoate were measured using an internal standard (n-butanol) GC headspace method and a flame ionization detector (FID). The carrier gas was helium at 6.0 mL/min and the GC was equipped with a Hewlett Packard 7694 headspace autosampler.
GC injector temperature was 200° C. and detector temperature was 220° C. Oven temperature profile: 40° C. (five minutes), 40-200° C. (10° C./min), 200-220° C. (50° C./min), 220° C. (five minutes). The FID gases included the carrier at 6.0 mL/min, helium makeup at 30 mL/min and 28 psig, H2 at 50 mL/min and 25 psig, and air at 300 mL/min and 35 psig.
The septum was purged at a flow rate of 0.8 mL /min. The head pressure was 4.0 psig. When the autosampler was connected via a needle in the injection port, the vial pressure was 15.9 psig, the carrier pressure was 7.1 psig, the column head pressure was 4 psig, the split flow was 18 mL/min and the column flow was 6 mL/min. Zone temperatures: vial at 70° C., loop at 80° C., transfer line at 150° C.
The GC cycle time was 40 minutes, with a vial equilibration time of 35 minutes, a pressurization time of 0.25 minute, a loop fill time of 0.1 minute, a loop equilibration time of 0.1 minute, an injection time of three minutes and a sample loop volume of 1 mL.
Free Amino Nitrogen (FAN)
The Free Amino Nitrogen International Method of the Technical Committee and Editorial Committee of the American Society of Brewing Chemists (1992) was used to determine the concentration of free amino nitrogen in a fermentation sample, using a spectrophotometer. This spectrophotometric method displays a color reaction between ninhydrin and the nitrogen present in the sample. The amount of absorbance is directly related to the amount of free amino nitrogen present.
The samples were diluted to a ratio of 100:1 with distilled water and 2 mL of the diluted sample were introduced into each of three test tubes. The blank was prepared by introducing 2 mL of distilled deionized water into each of three test tubes. Three test tubes containing 2 mL each of the glycine standard solution were also prepared.
For all samples, 1 mL of color reagent was added, and then they were placed in a 100° C. water bath for exactly 16 minutes. The test tubes were then cooled in a 20° C. water bath for 20 minutes. Five mL of the dilution reagent was then added to each test tube and mixed thoroughly. The samples were then allowed to stand for 10 to 15 minutes. The absorbance at 570 nm was then measured using a spectrophotometer and the amount of FAN in a sample was calculated using the following equation:
FAN (mg/L)=(AP−AB−AF)2d /AS
where FAN is the amount of free amino nitrogen in the sample in mg/L, AP is the average of the absorbances of the test solutions, AB is the average of the absorbances for the blanks, AF is the average of the absorbances for the correction for dark worts and beers, 2 is the amount of FAN in the glycine standard solution, d is the dilution factor of the sample, and AS is the average of the absorbances for the glycine standard solution.
Continuous Fermentation Using A Gas-lift Bioreactor System
A gas-lift draft tube bioreactor system was chosen for continuous beer fermentation because of its published excellent mass transfer (liquid-solid) and mixing characteristics. Liquid-solid mass transfer is especially important since it involves the transfer of nutrients from the liquid phase to the solid immobilized cell biocatalyst, providing substrates for the encapsulated yeast. These bioreactors also provide good aeration, low power consumption, and are simple to construct. This has made gas-lift bioreactor systems very attractive for large scale operations, such as those used commercially for wastewater treatment (Driessen et al., 1997; Heijnen, 1993).
Bioreactor Body
The 13 L (8 L working volume) gas-lift draft tube bioreactor designed for this work was a three phase fluidized bed (liquid/solid/gas) where the immobilized cells were kept in suspension by carbon dioxide gas driven internal liquid circulation (Heijnen, 1996) as shown in
A draft tube, concentrically located inside the columnar bioreactor, functioned as the riser in this fluidized bed system while the outside annulus served as the downcomer. The internal draft tube was suspended from a cylindrical particle separator, seated on three stainless steel tabs in the expanded head region of the bioreactor. Keeping the draft tube and particle separator fittings inside the bioreactor, minimized the risk of microbial contamination from the outside environment.
Originally, the bioreactor had a mesh screen to separate the immobilized cells from the liquid at the outlet. However, the screen was prone to plugging, so a stainless steel cylinder was used to separate the immobilized cell beads from the liquid phase as they moved over the top of the draft tube and flowed down the annulus. The particles would hit the cylinder and fall back down into the bulk liquid phase rather than leaving the bioreactor as overflow. Thus there was a small region near the bioreactor outlet that was free of immobilized cell particles. The bioreactor expanded head region also increased the surface area for gas bubble disengagement.
Bioreactor Headplate
In
Sanitary Valves for Aseptic Sampling
The bioreactor was equipped with a membrane sample valve welded into the bioreactor wall. The valve was designed for sampling under aseptic conditions. The membrane sealed directly against the fermentation liquid, allowing the valve to be fully sterilizable with steam and alcohol through two outlets (
A pulse tracer method was used to determine mixing time in the gas-lift bioreactor (Chistie, 1989). A 1 mL volume of 10 N HCI was rapidly injected into the bioreactor annulus, and the change in pH was logged over time, with time, t=0 seconds at the time of the injection. The pH was returned to its original value by injecting 10 N NaOH. The pH electrode was 277 mm in length and 3.5 mm in diameter. An Ingold Model 2300 Process pH Transmitter was used to monitor pH. A two-point pH calibration was performed with certified standard buffers. The data were logged at a frequency of 3750 Hz for 300 seconds using a software program.
The pH data were then smoothed using an algorithm. The algorithm is a time-domain method of smoothing based on least squares quartic polynomial fitting across a moving window within the pH data (Anon, 1996). The smoothed data were then normalized, and a plot of ΔpH versus time was generated. The mixing time was taken to the nearest minute, when the pH had reached ˜95 percent of equilibrium value. The mixing time was measured using three different volumetric flow rates of carbon dioxide: 283 cm3/min, 472 cm3/min (volumetric flow rate used throughout this work), and 661 cm3/min. In all three cases the pH in the bioreactor had equilibrated (˜95 percent cutoff) in less than two minutes. The mixing time was deemed to be sufficiently short to validate the inventors' original assumption that the bioreactor was well-mixed. This allowed the inventors to assume that the composition of the liquid sampled from the bioreactor wall was not significantly different from that which flowed from the outlet, with an average liquid residence time of 24 hours in the bioreactor.
Flow Diagram of Continuous Beer Fermentation System
A flow diagram for the continuous beer fermentation system is given in
Wort Collection and Storage
Unoxygenated wort for continuous fermentation was collected via piping into a 1600 L cylindroconical storage tank, pre-purged with carbon dioxide to minimize oxygen pickup by the wort. All tanks of this scale, including wort holding tanks, T-1 and T-2, were cleaned and sanitized prior to their use. The wort was then flash pasteurized and transferred at 2° C. into the available wort holding tank, T-1 or T-2 (also pre-purged with carbon dioxide). Wort was held in these tanks at 2° C. for up to two weeks, supplying liquid to the continuously fermenting bioreactor, BR-1. At the end of the two-week period, the bioreactor feed was changed over so that wort was supplied from the second wort tank, which contained fresh wort. Two identical wort storage tanks, T-1 and T-2, were employed to minimize downtime during the changeover to fresh wort. In all cases, wort was tested for contamination a minimum of two days prior to being introduced into the bioreactor (BR-1). If the wort was contaminated, it was discarded and fresh wort was immediately collected and pasteurized.
Minimizing Wort Dissolved Oxygen Concentration During Storage: The goal was to store the wort with minimal oxygen, at a constant level, and at a low temperature, without freezing the wort. This was required to prevent undesirable staling reactions in the wort from chemical reactions with oxygen (Narziβ et al., 1993), to provide a consistent supply of wort to the bioreactor, and to minimize the risk of wort contamination with microbes during storage. The large 1,600 L (net) cylindroconical vessels (T-1 and T-2) used to store the wort for the continuous fermentations were originally designed as batch fermenters, not wort storage tanks. Because of this, the cooling for these vessels was not adequate to maintain wort at 2° C. After three days of holding the wort, temperature varied as much as 15° C. from one region of the tank to another (Table 11).
These warm regions in the tanks increased the risk of microbial growth. Thus, some agitation was needed in these tanks to ensure a uniform low temperature throughout.
For these reasons, a pipe sparger was installed into the base cone of each wort storage tank (T-1 and T-2). Experiments were performed to determine the best protocol for filling the tank with wort and maintaining constant low levels of dissolved oxygen. In the first experiment, the storage tank was filled with wort that had been collected without oxygenation and flash-pasteurized. Once the storage tank was filled with 1,600 L wort, 0.113 m3/hr of carbon dioxide was sparged into the base of the tank. During the second experiment, the wort was again collected without oxygenation and flash-pasteurized. This time, the storage tank was purged with carbon dioxide (0.85 m3/hr) for three hours prior to filling, and a small amount of carbon dioxide (0.113 m3/hr) was continuously sparged into the storage vessel as the wort was being transferred into the tank. This low flow of carbon dioxide was continuously bubbled through the wort stored in the tank while it supplied wort to the continuous fermentation. For both experiments, wort dissolved oxygen concentration was monitored on a regular basis during a week of storage.
In
The temperature profile in the storage vessels was also compared with and without 0.113 m3/hr of carbon dioxide sparging. This was performed on water rather than on wort using a Type T temperature probe connected to a thermometer. City water (1,600 L) was collected into a wort storage tank and equilibrated for three days and the temperature of the water was recorded in different regions of the storage tank. The water in the tank was then sparged for 24 hours with 0.113 m3/hr carbon dioxide, and the temperature was again recorded. Ambient temperature was recorded in each case, and the temperature set point within the storage tank was 2.0° C.
As seen in Table 11, with carbon dioxide sparging, the temperature in the storage tanks was more uniform, with temperature ranging between 0.1 and 4.1° C. in the regions measured, and the contents of the tanks did not freeze. This lower temperature helped to prevent unwanted growth of microbes in the wort during storage.
Gas was released from the wort storage tanks through a sterile gas filter situated at the top of the tank. Wort was then transferred using a variable speed peristaltic pump to the 8 L bioreactor (BR-1) inlet using food grade L/S 16 flexible tubing.
Continuous Fermentation using Gas-Lift Draft Tube Bioreactor System
Wort was introduced near the bottom cone of the bioreactor, BR-1, through a ¼″ port. A mixture of filter-sterilized air and carbon dioxide (99.99 percent purity) flowed into the bioreactor through the sintered stainless steel sparger. A rotameter (R-3) was used to control the carbon dioxide flow rate at STP, and a precalibrated mass flow controller (M-1) was used to control the flow rate of air at STP. Fermented liquid left the bioreactor as overflow and flowed through 1″ I.D. reinforced PVC tubing into a 30 L stainless steel collection vessel (T-3) which was cooled with an external glycol coil and kept at a temperature of 4° C.
Product Collection
The product collection vessel (T-3) had a large inlet port (1″ I.D.) which was designed so that the fermented liquid would flow down the collection vessel wall to minimize foaming. This vessel also had a sterile gas filter for gas release from the bioreactor (BR-1) and the collection vessel (T-3). The collection vessel was periodically emptied using a ¼″ valve (V-12) situated 2″ above the base of the tank.
Glycol Cooling Loop
Glycol was transferred at a temperature of −23° C. and pressure of 45 psig, and circulated through cooling jackets for the wort holding tanks (T-1 and T-2), the gas-lift bioreactor (BR-1), and the product collection vessel (T-3). The two wort holding tanks and the bioreactor were equipped with liquid phase temperature probes which provided feedback to temperature controllers, which in turn controlled the flow of cold glycol to the vessel jackets. The wort holding tanks stored the wort at 2° C., while the temperature within the bioreactor was controlled at temperatures of 12° C. to 22° C., depending upon the specific experiment. The product collection vessel did not have automatic temperature control, but rather, the flow of glycol was manually controlled to keep the vessel at approximately 4° C. It was not necessary to control precisely the temperature of the product collection vessel (T-3) because the liquid in this vessel was simply discarded and not analyzed or processed further.
Glycol was also used to jacket and cool the wort transfer lines from the wort tanks (T-1 and T-2) to the bioreactor (BR-1). Once the glycol had circulated through a given jacket, it was returned to a main line and then was returned plant, generally at a temperature of −15° C. and pressure of 40 psig.
Bioreactor Sterilization Protocol
The bioreactor (BR-1) was filled with a two percent (v/v) solution of a sanitizing detergent, and soaked overnight with gas sparging. The reactor was then drained and rinsed with cold water. This cycle of cleaning solution and water rinsing was repeated two times. In order to prepare the bioreactor for steam sterilization, the wort and gas lines were disconnected. The steam line was connected to the bioreactor inlet and the following valves were opened: the bioreactor inlet and purge valves (V-7, V-6), the gas inlet (V-17), product outlet valves (V-9, V-11), the membrane sampling valves (V-8, V-10), and collection vessel drain port (V12). The plant steam valve was then slowly opened, and the bioreactor valves were adjusted so that a trickle of steam was observed at the exit of each external opening. After 60 minutes of steam exposure, all the external valves on the bioreactor were closed (V-17, V-8, V-10, V-12) except the wort bypass valve (V-6). When the steam valve was closed, the wort bypass valve was closed, and a sterile filter was connected to the collection vessel to prevent contamination by non-sterile air entering the system as it cooled. The bioreactor gas line was also reconnected at V-17 as the plant steam line was closed in order to maintain a positive pressure while the system cooled.
Fermentation System Startup
Brewer's wort was collected from the plant into a 20 L stainless steel pressure vessel and heated in an autoclave for 45 minutes at 100° C. Immobilized cells were aseptically transferred into the cooled wort (40 percent v/v). The sealed vessel was transported to the Microbrewery Pilot Plant where the bioreactor system was housed. The 20 L vessel was connected to a quick connect fitting, which was clamped to reinforced ⅜″ PVC tubing. The other end of the PVC tubing was clamped to the membrane sampling valve (V-8) in the bioreactor wall. Filter-sterilized carbon dioxide was applied as 10 psig to the 20 L vessel, and the membrane sample port was opened so that the immobilized cell mixture was transferred from the vessel into the bioreactor, without exposing the inoculum to the outside air environment. The internal components of the “quick connect” fittings of the 20 L vessel were removed to prevent plugging with immobilized cells upon transfer into the bioreactor. The cumulative particle size distribution (undersize) for the kappa-carrageenan gel beads is shown in
Following inoculation with immobilized cells, the bioreactor was operated in batch mode until the sugar and diacetyl concentrations reached targets of less than 3° Plato in terms of specific gravity and less than 100 μg/L diacetyl. The system was then prepared for continuous operation. In order to rinse with hot water and steam-sterilize the wort transfer line, valves V-2 (or V-4 for T-2), V-5, and V-6 were opened, while V-1 (or V-3 for T-2) and V-7 were closed, isolating the wort line. The wort transfer line was rinsed with hot water at approximately 80° C., which was supplied through V-2 (or V-4 for T-2). Following the hot water rinse cycle, the plant steam line was connected at the same location, and the wort transfer line was steam sterilized for a minimum of 30 minutes. At the same time that the steam line was shut off, the bypass valve (V-6) was also closed. Once the system had cooled, V-2 (or V-4 for T-2) was closed, and the steam line was disconnected. The wort tank valve, V-1 (or V-3 for T-2), and the bypass valve (V-6) were opened, and the wort transfer pump (P-1) was started. The wort was sent to the sewer drain via the bypass valve (V-6) until the condensate in the line was replaced with fresh cold wort. At that point the bypass valve was closed and the bioreactor inlet valve (V-6) on the reactor was opened, commencing the continuous fermentation process.
Every two weeks the tank, which supplied wort, was alternated between storage tanks (T-1 and T-2). After two weeks of supplying wort from T-1, the continuous feed pump, P-1, was stopped and the valve (V-5) at the inlet of the bioreactor was closed. The wort transfer line was then connected to the second storage tank (T-2) and the line was flushed and sterilized as described in the previous paragraph. Continuous fermentation then resumed after only a short down time of less than one hour.
Kappa-Carrageenan Gel Immobilization of Lager Brewing Yeast
Scientists have studied a variety of matrices for the physical entrapment of whole cells including calcium alginate (Bejar et al., 1992; Curin et al., 1987; Masschelein and Ramos-Jeunehomme, 1985; Nedovic et al., 1996; Shindo et al., 1994; White and Portno, 1978), agarose (Hooijmans et al., 1990; Lundberg and Kuchel, 1997), and carrageenan gels (Norton et al., 1995; Wang et al., 1982). Carrageenan is a food grade material and it has been favored for cell encapsulation due to its superior mechanical strength compared to other gels (Büyükgüngör, 1992).
The yeast cell colonization within kappa-carrageenan gel beads was monitored over three cycles of repeated batch fermentation. The viability of the immobilized cells and the cells released into the liquid phase was examined. Fermentation parameters including ethanol, maltose, maltotriose, fructose, and glucose were followed throughout the repeated batch fermentations and then compared with control fermentations using only freely suspended yeast cells under the same nutrient conditions.
There has been little published information to date on the physical effects on cells after long term immobilization (Virkajärvi and Kronlöf, 1998) and continuous exposure to external stresses and fermentation products. The viability, cell population distribution and physical appearance of yeast cells immobilized within carrageenan gel beads over an extended period of continuous fermentation in a gas-lift bioreactor was also examined. Also evaluated, over extended periods of time, were the relative percentages of respiratory deficient yeast in the immobilized and freely suspended cell population of the bioreactor.
Carrageenan is made up of repeating 3-6-anhydrogalactose units and assorted carrageenans differ by the number and position of the sulfate ester groups on repeating galactose units. A schematic of the carrageenan gelation mechanism may be seen in
The three common types of carrageenan are lambda, iota, and kappa As illustrated in
An important characteristic of carrageenan is its reversible thermogelation properties. As carrageenan solution is cooled, viscosity increases and gelation occurs. As the solution is heated, viscosity decreases and the carrageenan reverts to the sol state. By controlling the composition of the gelling cation solution, the temperature at which carrageenan is transformed from a sol into a gel may be altered. Kappa-carrageenan gelling temperature increases with increasing potassium chloride concentration in solution. This phenomenon was used to engineer a process for cell immobilization, since severe temperature fluctuations can be avoided (Neufeld et al., 1996). The gelling temperature of the carrageenan can be controlled such that it is high enough to be a gel under fermentation conditions, yet low enough that the yeast cells may be mixed with the carrageenan in its sol state without detrimental effects on viability prior to bead gelification.
There are a number of factors, nevertheless, which indicate the need for further study on the effects of immobilization within gel matrices on yeast cell metabolism and physiology. Immobilized cells are not subjected to the same micro-environment as the free cells in the liquid phase because there are additional barriers from the gel matrix and other entrapped yeast cells which must be surmounted, before substrates can be transported to their surfaces (
The age of the immobilized yeast must also be considered. Entrapped cells age as a continuous fermentation proceeds over the course of months and they ferment under a defined set of pseudo-steady-state conditions. However, during batch fermentation, yeast cells are exposed to an environment that changes with time and the cells are only reused for a limited number of fermentations before disposal. More research is needed to study the long-standing effects of continuous fermentation on yeast cell vitality, relating to fermentation performance.
Kappa-Carrageenan Gel Bead Production: Kappa-carrageenan gel beads containing entrapped lager yeast cells were produced using the static mixer process with an initial cell loading of 2.6×107 cells/mL of gel (Patent Application 2133789 (Neufeld et al., 1996) and a bead diameter of 0.5 to 2.0 mm.
Fermentation Medium: Brewery wort with a specific gravity of 17.5° P as described in detail previously.
Repeated Batch Kinetics of Yeast Immobilized in Kappa-Carrageenan Gel Beads
Fermentations were conducted in 2 L Erlenmeyer flasks at 21° C., with shaking at 150 rpm. Carrier loading was 40 percent (v/v) of immobilized cell beads and the total fermentation volume was 1 L. Each fermentation was seven days in duration. In R1 fresh immobilized cell beads were pitched into wort and at the end of the fermentation these beads were separated from the fermented liquid by passing the mixture through a sterile stainless steel sieve (500 μm mesh size). The beads were then repitched at the same proportion into fresh, sterile wort for a second (R2) and then third (R3) batch fermentation. Sampling was performed twice a day for the first three days, and then once per day for the fourth and fifth days of each fermentation. Fermentations were carried out in duplicate or triplicate. All fermentations were conducted with freely suspended cell control fermentations, which were conducted under the same conditions except that only free cells were pitched into the fermentations at a rate of 4 g/L. Samples were analyzed for free and immobilized cell viability and cell concentration, and liquid phase carbohydrate and ethanol concentrations.
Yield factors, YP/S, of product ethanol, from substrate total fermentable glucose, were calculated using equation 3.20 for the three immobilized cell fermentation cycles and the free cell control. For all fermentations the yield factors were calculated from the start of fermentation to the time that maltose consumption was complete.
Ethanol productivity, Vethanol, the amount of ethanol produced per total bioreactor working volume per unit fermentation time was calculated using equation 3.25 for R1, R2 and R3, and the free cell control from the start of fermentation to the time that maltose consumption was complete. In the case of the yield factors and ethanol productivity, the contributions of the immobilized and freely suspended yeast cells were not distinguished from one another.
The local maximum specific growth rate, μmax, and cell doubling time were calculated for the averaged free cell control using equations 3.3 and 3.4.
Viability and Morphological Characteristics of Immobilized Yeast over Extended Fermentation Time
Batch Fermentation Conditions: Batch fermentations were conducted in 2 L Erlenmeyer flasks at 21° C., with shaking at 150 rpm. Carrier loading was 40 percent (v/v) with a total fermentation volume of 1 L.
Continuous Fermentation Conditions: Pilot scale gas lift draft tube bioreactors were used for continuous fermentations. All data collected were from an 8 L working volume bioreactor, except the two month scanning electron micrographs, which were collected from a 50 L bioreactor using the same fermentation medium and immobilization method. Immobilized cell beads at 40 percent (v/v) were fluidized within the bioreactors using a mixture of air and carbon dioxide. The bioreactors were operated under varying conditions with fermentation temperatures controlled at 12, 17, and 22° C. and residence times held between 0.9 and 1.8 days. The gas lift reactor reached a maximum ethanol concentration of 73 kg/m3 during the six-month experiment, with an average concentration of 58 kg/m3.
Microbiological Analyses: Samples were taken from the liquid phase of the gas lift bioreactor at least once a week to test for contaminants including wild yeast, non-lager yeast, and aerobic and anaerobic beer spoilage bacteria. After five months, the liquid phase yeast cells were assayed in duplicate for respiratory deficient mutation.
Scanning Electron Microscopy (SEM): Kappa-carrageenan gel beads (1.0-1.5 mm diameter) containing immobilized lager yeast cells were sampled for SEM examination at four different times: (1) after immobilized cell bead production and before inoculation of beads into the fermentation medium; (2) after two days in batch fermentation; (3) after two months of continuous fermentation in a pilot scale gas lift draft tube bioreactor; (4) after six months of continuous fermentation in a pilot scale gas lift draft tube bioreactor. The methodology used for SEMs and the related sample preparation are described in section 4.7. Using the methods described in section 4.6, yeast cell concentration and viability (immobilized and freely suspended) were assessed at the same times as the SEMs.
Repeated Batch Kinetics of Yeast Immobilized in Kappa-Carrageenan Gel Beads
Fermentation time was greatly reduced each time the immobilized cells were repitched into fresh wort, as seen in
In R1, the immobilized cell concentration within the kappa-carrageenan gel bead was increasing at a similar rate to the control fermentation, which contained only liquid phase cells. This was confirmed by comparing the average growth curve of the free cell control fermentations in
In Table 12, the yields of ethanol from substrate fermentable sugars, YP/S, are shown for the three batch generations and the control. In Table 13, the bioreactor volumetric productivities of ethanol are also given, calculated using the data given in Table 12. The yields of ethanol from sugars for the fermentations were not significantly different from each other or from the control. Yields were all above 90 percent of the theoretical yield of 0.51 predicted from the Guy-Lussac equation. As mentioned earlier, biomass production and other by-products formed by the yeast cells prevent efficiencies from reaching higher than 95 percent of theoretical (Hardwick, 1995). The volumetric bioreactor productivity of ethanol in the three repeated batch fermentations varied significantly from batch to repeated batch. Ethanol productivity increased with each cycle of repeated batch fermentation and, by R3, the immobilized cells were more productive than the control fermentation. The total amount of ethanol produced in R2 was not significantly greater than that produced during R1, but the fermentation time was less than half of R1 and of the control fermentation.
There are many factors that could contribute to this increased fermentation rate of immobilized cells with each batch repetition, such as yeast cell adaptation to the fermentation conditions and the progressively increasing cell concentration. The total number of cells per bioreactor volume only becomes significantly greater than that of the control by R3. In
*The symbol tf is the time in hours to complete maltose uptake and the subscript f refers to the concentration of the given analyte at time, t = tf.
*calculated when maltose uptake complete.
Another factor affecting the increased bioreactor volumetric productivity observed with each repeated batch fermentation, involves yeast cell adaptation. By the end of the first fermentation, yeast cells had adapted their metabolic machinery to the given fermentation conditions. This may result in a decrease in the lag phase at the beginning of subsequent batch fermentations, increasing the rate of fermentation. During this study all control fermentations were carried out with freshly prepared lager yeast. It would be interesting to repitch the freely suspended control yeast alongside the repitched immobilized cells to further examine this effect relative to the cell concentration effects.
Using the averaged data from three separate freely suspended yeast control fermentations, a plot of ln (X/Xo) versus fermentation time is given in
Viability and Morphological Characteristics of Immobilized Yeast over Extended Fermentation Time
Before the gel beads were exposed to fermentation medium, and following immobilized cell bead production using the static mixer process, the cell concentration was 2.6×107 cells/mL of gel bead (Table 14, where values are the averages of two samples). SEM imaging shows the cells to be individually and uniformly distributed throughout the gel bead (
*Based on single sample.
Viability was >90 percent following two days of batch fermentation, and cell concentration within the gel bead had increased ten-fold (Table 14). Cells (>90 percent viable) had also begun to be released from the gel into the bulk liquid phase of the fermentation, yielding a concentration of 107 cells/mL of liquid. Small yeast colonies formed inside the gel beads, with many bud scars present on individual cells as seen in
Immobilized yeast cell viability decreased after two months of continuous fermentation in a gas lift bioreactor (Table 14), but the cells in the bulk liquid phase remained highly viable (>90 percent), and this finding was supported during several different continuous fermentations in pilot scale gas lift bioreactors. The SEM in
The viability of the yeast immobilized within the carrageenan gel had dropped to below 50 percent after six months of continuous fermentation in the gas lift bioreactor, (Table 14). It should be noted that while only a single data point for immobilized cell concentration and viability was collected at six months, data at the five-month mark were similar, with an immobilized cell concentration of 1.14×109 cells/mL of gel and viability of <50 percent. While a gradual decline in immobilized cell viability was seen over time, the viability of the cells in the bulk liquid phase remained reliably high. In addition, even though immobilized cell viabilities were low in the beads as a whole, the bioreactor produced a fully fermented beer during its sixth month of continuous operation. Possible reasons for this finding include the significant contribution of the highly viable freely suspended yeast cells to the fermentation. As well, there is the potential contribution of viable immobilized cells located at the periphery of the gel bead where there are fewer barriers to mass transfer, as compared to the cells located at the center of the bead. It is unclear whether the immobilized cells had the ability to redistribute themselves within the gel matrix, or if these cells remained stationary where they were first located. A concentration of 109 cells/mL of gel bead was the maximum reached within these beads over the six-month period of continuous fermentation.
In
Over the entire six month beer fermentation trial, the gas lift bioreactor was tested a minimum of once a week for contamination. No bacterial contaminants were detected at any time during the trial. In the last two months of the trial, a contaminating yeast was detected in concentrations which fluctuated between one and five cfu/mL. This yeast was capable of growth on PYN medium at 37° C. but did not grow aerobically or anaerobically on DUBA medium (selective for bacteria), did not ferment dextrins, and showed no growth on CuSO4 medium (selective for wild yeast).
After five months, the average percentage of respiratory deficient yeast cells was seven percent, which is higher than what is normally found using this strain during industrial batch fermentations (two percent average). Other researchers have reported similar findings (Norton and D'Amore, 1995). Respiratory deficient yeasts result from a mutation which causes yeast to be incapable of respiring glucose to carbon dioxide and water. These yeasts have mitochondria with permanently impaired activity and arise usually because of a mutation of mitochondrial DNA (Hardwick, 1995).
Artifacts from SEM sample preparation can cause confusion. Technologies such as nuclear magnetic resonance (NMR) spectroscopy (Fernandez, 1996) and confocal microscopy (Bancel and Hu, 1996) have been used to examine immobilized cells non-invasively. NMR imaging techniques have allowed researchers to study transport, flow and spatial distribution of cells and biochemicals in biofilms. Researchers (Bancel and Hu, 1996) have also shown that confocal laser scanning microscopy can be used to observe cells immobilized in porous gelatin microcarriers through serial optical sectioning.
Although methylene blue is used as a standard indicator of cell viability in the brewing industry the method has many shortcomings (Mochaba et al., 1998). It measures whether a yeast population is viable or non-viable based on the ability of viable cells to oxidize the dye to its colorless form. Non-viable cells lack the ability to oxidize the stain and therefore remain blue (O'Connor-Cox et al., 1997). Plate count and slide culture techniques are based on the ability of the cells to grow and produce macrocolonies on agar plates or microcolonies on media covered microscope slides (Technical Committee and Editorial Committee of the ASBC, 1992). Ongoing work of examining the viability of yeast in immobilized matrices over extended periods of time at Labatt now uses not only methylene blue but also the aforementioned methods as well as developing the confocal microscopy technique using vital staining. In addition to measuring the viability of the cells, the issue of “vitality” of the immobilized cells must also be addressed in future work. Where viability has been used to describe the ability of cells to grow and reproduce, vitality measures yeast fermentation performance, activity, or the ability of the yeast to recover from stress (Smart et al., 1999).
Flavor Production in a Gas-Lift Continuous Beer Fermentation System
Experimental Procedure
Using continuous fermentation to produce beer is very different from other applications using immobilized cells because the resulting product is not measured in terms of one component of interest such as ethanol. Rather, it is a balance of numerous chemical compounds which must be balanced to make a quality finished product. The effects of oxygen on yeast flavor metabolites during continuous primary fermentation and during a secondary batch holding period were examined. The effect of residence time on flavor metabolites was also examined at two levels. Lastly, a commercial enzyme preparation of alpha-acetolactate decarboxylase was added to the continuous fermentation wort supply, and liquid phase total diacetyl concentration was monitored.
Effect of Relative Amounts of Air in the Bioreactor Fluidizing Gas on Yeast Metabolites During Primary Continuous Fermentation
The amount of air and hence oxygen in the bioreactor fluidizing gas was varied while residence time, temperature, and all other controllable process variables were held constant. The total volumetric flow rate of gas was held constant at 472 mL/min at STP, temperature was 15° C., and kappa-carrageenan gel beads containing immobilized LCC 3021 yeast were used throughout the trial with an initial cell loading of 1×108 cells/mL of gel. Four different volumetric flow rates of air were imposed on the system throughout the trial (Table 15), and the average bioreactor residence time, Rt, was 1.18 days.
The following analyses were performed repeatedly throughout the experiment: free amino nitrogen (FAN), total fermentable carbohydrate (as glucose), ethanol, total diacetyl, beer volatiles (selected esters and alcohols), and liquid phase yeast cell concentration and viability. The bioreactor was also tested for contamination a minimum of once a week.
The dissolved oxygen concentration in the bulk liquid phase of the bioreactor was measured when the continuous fermentation was assumed to be at pseudo-steady state for each volumetric flow rate of air (minimum of three reactor turn-over times).
Post Fermentation Batch Holding Period: Effects of Oxygen Exposure on Yeast Metabolites
Even when the amount of oxygen in the bioreactor fluidizing gas was relatively low (34 mL/min at STP), the concentrations of acetaldehyde and total diacetyl found during the experiment conducted in section 7.1.2, were unacceptable high for the North American lager beer market. Therefore a novel approach was taken, where liquid taken from the continuous primary bioreactor was held in batch for 48 hours at a slightly elevated temperature of 21° C. to reduce the concentration of these two compounds. As well, the results from the previous section 7.1.2 indicated the significant effect that the amount of air in the fluidizing gas had on the flavor compounds measured. Therefore the effect on yeast flavor metabolites of aerobic versus anaerobic conditions downstream of the primary fermentation, where the secondary batch hold occurred, was examined.
Continuous primary fermentation was performed in a 50 L gas lift bioreactor using a highly flocculent variant of the LCC3021 yeast strain for this trial because the sample volume requirement for the study was too large relative to the volume of the 8 L bioreactor. Operating conditions were 1180 mL/min CO2 and 189 mL/min air at STP in the fluidizing gas, an average bioreactor residence time, Rt, of 1.0 day, a temperature of 15° C., and high-gravity 17.5° P lager brewer's wort.
A total of four samples were taken (100 mL crimp vials), with two handled under anaerobic conditions and the other two were exposed to the aerobic environment.
The anaerobic sampling procedure was as follows: two 100 mL crimp vials and six 25 mL crimp vials were autoclaved and then placed in an anaerobic box (Labmaster 100, mbraun, USA) with argon as the purging gas. The 100 mL vials were allowed to equilibrate for 45 minutes and then they were sealed using aluminum caps and TEFLON® septa. A 50 mL syringe, fitted with a 3 inch, 16 gauge needle, sanitized using a 70 percent (v/v) ethanol solution, was used to withdraw sample from the bioreactor by puncturing the septum of the membrane of the sample valve and the sample was injected into the 100 mL pre-purged anaerobic vials. It was necessary to provide a vent to the crimp vial, through an additional sterile syringe needle, to allow release of the pressure within the vial during filling. The aerobic samples were exposed to the atmosphere as they were drained from the bioreactor by fully opening the membrane sample valve into the 100 mL unsealed sample vials, without using a syringe and needle.
The sample liquid was allowed to rest at room temperature for two hours in order to allow the yeast to settle out of solution, leaving a cell concentration in the bulk liquid of approximately 106 cells/mL. Once settled, the liquid from each 100 mL vial was decanted into three 25 mL vials. The anaerobic samples were handled in an anaerobic box in order to minimize oxygen pickup while the aerobic samples were processed under the laminar flow hood. Each of the samples in the 100 mL vials was split into three smaller 25 mL vials, so that sample analyses could be performed without altering the course of the fermentation due to sample removal. Once the aerobic samples were transferred to the smaller vials they were incubated, uncapped at 21° C. The anaerobic samples were transferred to the three smaller vials and sealed using an aluminum cap and TEFLON® septa. In order to avoid pressure buildup due to carbon dioxide evolution within the vials, while preventing exposure of the samples to the aerobic external environment, the septa were punctured with a needle. The end of the needle exposed to the external environment was submerged in ethanol (less than 1 cm of pressure head), preventing any back-flow of air into the sample. Samples were collected for analysis at 2, 24, and 48 hours. A sample was also taken directly from the bioreactor and analyzed immediately in order to assess the state of the fermentation within the bioreactor at the time of the protocol. The samples were analyzed for total fermentable carbohydrate (as glucose), ethanol, total diacetyl, and beer volatiles (selected esters and alcohols).
Effect of Liquid Residence Time on Yeast Metabolites During Continuous Primary Beer Fermentation
In order to examine the effect of liquid residence time on yeast metabolic activity, a trial was performed in which a step change in wort volumetric flow rate to the bioreactor was imposed during continuous primary beer fermentation using LCC3021 yeast cells immobilized in kappa-carrageenan gel beads. The bioreactor temperature was held constant at 17° C. throughout the trial. The gas volumetric flow rate supplied to the bioreactor was also constant at 472 mL/min at STP. The gas was a mixture of air (11 mL/min at STP) and carbon dioxide (461 mL/min at STP). The initial concentration of yeast cells in the kappa-carrageenan gel was 2.6×107 cells/mL of gel bead and the bioreactor contained 40 percent (v/v) of beads.
The following analyses were performed repeatedly throughout the trial: carbohydrates, free amino nitrogen (FAN), total fermentable carbohydrate (as glucose), ethanol, total diacetyl, beer volatiles (selected esters and alcohols), and liquid phase yeast cell concentration and viability. The bioreactor was also tested for contamination, a minimum of once a week.
Using a Commercial Preparation of Alpha-Acetolactate Decarboxylase to Reduce Total Diacetyl During Continuous Primary Beer Fermentation
High diacetyl concentrations are considered by most North American brewers to be an undesirable flavor defect in their beer. In the continuous primary fermentations performed to date, total diacetyl concentrations have consistently been above the threshold levels for traditional batch fermentations (70-150 μg/L) in a North American lager. During batch fermentation, diacetyl is reduced during the later stages of fermentation, when oxygen is no longer present and additional sugars are not being introduced. In the continuous fermentation system a constant low level of oxygen is supplied to the bioreactor through a sparger, and fresh wort is continuously supplied to the bioreactor. Therefore, a novel strategy using a commercial enzyme preparation was explored to control diacetyl concentration in the continuous bioreactor.
In a wort fermentation diacetyl is formed when alpha-acetolactate, an intermediate in the synthesis of valine, is oxidatively decarboxylated outside the yeast cell. The yeast cell then reabsorbs diacetyl and converts it into the less flavor-active acetoin. This oxidative decarboxylation of alpha-acetolactate to diacetyl is rate-limiting in batch wort fermentations. During the continuous fermentations, total diacetyl exited the bioreactor at unacceptably high concentrations (300-400 μg/L). The commercial enzyme alpha-acetolactate decarboxylase (ALDC), from Novo-Nordisk A/S can convert alpha-acetolactate directly into acetoin, thus avoiding the unwanted diacetyl intermediate (
Alpha-acetolactate decarboxylase was added to the wort fed into the bioreactor in order to examine its net effect on total diacetyl concentration. Other strategies for reducing diacetyl, including a batch warm hold period of 48 hours post-fermentation and immobilized secondary fermentation systems, technology from Alpha-Laval (Anon, 1997), were also explored. Both of these other strategies have shown success in reducing diacetyl levels post fermentation, but neither addresses the level of diacetyl at the source (i.e. at the bioreactor outlet). By using ALDC in the wort to reduce the diacetyl concentration coming out of the bioreactor, the post-fermentation treatment periods could be minimized or eliminated.
ALDC activity is optimal at pH 6.0 in lager wort at 10° C. At pH 5.0, typical of industrial worts, ALDC activity is maximized at a temperature of 35° C. (Anon, 1994). Thus under typical beer fermentation conditions of reduced temperature and pH, ALDC activity is less than optimal.
Health Canada in 1997 amended Canada's Food and Drug Regulations (SOR/97-81) to allow the use of ALDC in alcoholic beverages, which has opened the door for its use in Canadian breweries. Bacillus subtilis, carrying the gene coding for ALDC (E.C. 4.1.1.5) from Bacillus brevis, produces the enzyme ALDC. Because ALDC is an enzyme that is produced by a genetically modified organism (GMO), there are public perception issues that would need to be addressed before using such an enzyme in a commercial product.
Lager yeast, LCC3021, was used for these experiments as well as high gravity, 17.5° P, lager brewer's wort. Ethanol, total fermentable carbohydrate (as glucose), total diacetyl, and liquid phase cell concentration were monitored. Yeast cells were immobilized in kappa-carrageenan gel beads as described previously. The bioreactor was allowed three turnover times, before it was assumed to have reached pseudo-steady state. As mentioned earlier, the diacetyl method used in this work is referred to as “total diacetyl” because the method measures the amount of diacetyl and its precursor, alpha-acetolactate. Thus an observed reduction in total diacetyl during this experiment would be due to the combined effect of the enzyme converting alpha-acetolactate directly into acetoin and the subsequent lowered concentration of its derivative, diacetyl.
Alpha-acetolactate decarboxylase (ALDC) was supplied. The activity of the enzyme was 1500 ADU/g, where ADU is the amount of enzyme which under standard conditions, by decarboxylation of alpha-acetolactate, produces one μmole of acetoin per minute as described in Novo Nordisk Method AF27 (Anon, 1994).
Continuous Fermentation Conditions: Continuous fermentations were performed in the 8 L gas lift draft tube bioreactor pitched at 40 percent (v/v) with kappa-carrageenan gel beads containing immobilized lager yeast cells. The bioreactor was sparged with a mixture of carbon dioxide (438 mL/min at STP) and air (34 mL/min at STP). Fermentation temperature was controlled at 15° C. throughout the trials and the bioreactor residence time, Rt, was 1.5 days. Total diacetyl concentration was monitored under these conditions and an average pseudo-steady state control diacetyl concentration was reached. ALDC was then added to the wort at a concentration of 72 μg/L (108 ADU/L) and total diacetyl concentration in the bioreactor was monitored for a response.
Trial 1: Wort was collected from the brewhouse into a 20 L stainless steel vessel, and heated in an autoclave for 45 minutes at 100° C. The wort was held at 2° C. in a controlled temperature water bath while feeding the bioreactor. Once a pseudo-steady state total diacetyl concentration had been reached within the bioreactor, 72 μg/L (108 ADU/L) of ALDC was added to the wort inside the 20 L vessel. The initial biomass loading in the kappa-carrageenan gel beads was 3×107 cells/mL of gel.
Trial 2: In order to minimize the risk of contamination, the system was closed loop at the outlet and other upgrades were also made to the system as described in Chapter 4. As with Experiment 1, wort was collected from the brewhouse into a 20 L stainless steel vessel, and autoclaved for 45 minutes at 100° C. While feeding the bioreactor, the wort was held at 2° C. in a controlled temperature water bath. The initial biomass loading in the kappa-carrageenan gel beads was 3×107 cells/mL of gel. Once a pseudo-steady state total diacetyl concentration had been reached within the bioreactor, 72 μg/L (108 ADU/L) of ALDC was added to the wort inside the 20 L vessel.
Trial 3: Unoxygenated 17.5° P brewery wort (14 hL) was collected into a large wort storage vessel (T-1) in the Pilot Plant. It was then flash pasteurized and stored with carbon dioxide sparging in order to maintain a constant dissolved oxygen concentration of <0.10 mg/L. The wort was fed into the bioreactor from this tank until a pseudo-steady state total diacetyl concentration was reached. ALDC (72 μg/L) was then aseptically added to the wort for the remainder of the trial. The addition of ALDC was accomplished by measuring the amount of wort remaining in the storage vessel and calculating the amount of ALDC needed to bring the concentration of enzyme up to the target concentration of 72 μg/L (108 ADU/L). The appropriate amount of enzyme was then dissolved in 10 L of sterile wort. This solution was transferred to a 20 L stainless steel pressure vessel, which was connected via sterile tubing to the sample port on the wort holding vessel (T-1). The ALDC solution was then pushed using sterile carbon dioxide pressure into the wort holding vessel. In order to ensure that the ALDC solution was adequately mixed with the wort in the holding vessel, the flow rate of carbon dioxide sparged into the tank was increased to 4720 mL/min at STP for one hour and then returned to its normal flow rate. The storage tank then held enough ALDC dosed wort to complete the trial. The initial biomass loading in the kappa-carrageenan gel beads was 108 cells/mL of gel.
Effect of Relative Amounts of Air in the Bioreactor Fluidizing Gas on Yeast Metabolites During Primary Continuous Fermentation
In FIGS. 127 to 136 liquid phase yeast viability and cell concentration, free amino nitrogen (FAN), total fermentable carbohydrate (as glucose), ethanol, total diacetyl, acetaldehyde, ethyl acetate, 1-propanol, isobutanol, isoamyl acetate, isoamyl alcohol, ethyl hexanoate, and ethyl octanoate concentrations are plotted versus continuous fermentation time. All bioreactor operating conditions were held constant throughout the protocol except the percentage of air in the bioreactor sparging gas, which is marked directly on the figures. In Table 16 the averages for each analyte at pseudo-steady state (after a minimum of three reactor turnover times) are summarized.
*average of final four days of each operating condition
In
The ethanol concentration within the bioreactor liquid phase decreased, while total fermentable carbohydrate (as glucose) increased when the volumetric flow rate of air in the sparge gas was increased from 94 to 354 mL/min, as seen in
In
In
In Table 16 and
The higher alcohols isoamyl alcohol, isobutanol, and 1-propanol versus continuous fermentation time are given in
Because excess diacetyl, acetaldehyde and fusel alcohols are undesirable in beer, oxygen control to limit their production is important. As discussed in the literature review, when the supply of oxygen to the yeast cells is increased, there is enhanced anabolic formation of amino acid precursors and thus an overflow of higher alcohols, oxo-acids, and diacetyl. The concentration of esters is known to decrease with an increase in oxygen availability because ester formation is catalyzed by acetyl transferase. Acetyl transferase is inhibited by unsaturated fatty acids and ergosterol, which in turn will increase in the presence of oxygen (Norton and D'Amore, 1994).
For the bioreactor conditions used in this trial, the pseudo-steady-state (after a minimum of three reactor turnover times) dissolved oxygen concentrations measured in the liquid phase of the bioreactor were close to zero (less than 0.03 mg/L).
This trial did not allow for a direct comparison of the data from the 94 mL/min and 34 mL/min of air at STP in the fluidization gas, because they were separated by the highest air flow rate (354 mL/min). This is because the physiological state of the yeast resulting from the exposure to previous bioreactor conditions, the immobilization matrix and continuous fermentation time may also have caused other changes in flavor production.
No contamination was detected in the bioreactor at any point during this trial.
Post Fermentation Batch Holding Period: Effects of Oxygen Exposure on Yeast Metabolites
Because total diacetyl was not within normal ranges for a commercial beer at the end of primary fermentation, several approaches were taken to reduce the concentration of this compound to acceptable levels. One such approach was to use a warm holding period immediately following continuous primary fermentation.
In FIGS. 137 to 146 liquid phase total fermentable carbohydrate (as glucose), ethanol, total diacetyl, acetaldehyde, ethyl acetate, 1-propanol, isobutanol, isoamyl acetate, isoamyl alcohol, and ethyl hexanoate concentrations are plotted versus post fermentation holding time. Samples collected from the continuous primary fermenter at pseudo-steady state were held under aerobic or anaerobic conditions, as indicated in the legend of each figure.
In
The concentration of ethanol in
In
The concentration of total diacetyl versus holding time is given in
In FIGS. 141 to 143 the esters ethyl acetate, isoamyl acetate, and ethyl hexanoate concentrations are plotted versus post fermentation holding time. The same pattern for aerobic and anaerobic samples was observed for all esters. The concentration of esters did not diverge between the anaerobic and aerobic samples until later in the holding period, where the concentration of esters in the aerobic samples declined and the concentration in the anaerobic samples increased. Because the concentration of esters in the continuous fermentations is somewhat low compared with ester concentrations found in commercial beer, it is desirable to select conditions, which favor ester production.
FIGS. 144 to 146 show isoamyl alcohol, 1-propanol, and isobutanol concentration versus post fermentation holding time. At the end of the 48 hour holding period, no significant differences in these alcohols were observed between the aerobic and anaerobic treatments. However, the 24-hour samples showed a higher concentration in all cases for the aerobic treatments.
In
The formation of 1-propanol occurred during the continuous primary fermentation stage and it did not decrease significantly during the holding period, whether the conditions were aerobic or anaerobic. Kunze (1996) states that the following factors will increase higher alcohols such as 1-propanol during batch fermentation: mixing, intensive aeration of the wort, and repeated addition of fresh wort to existing yeast.
Ultimately the ideal scenario will be to eliminate the secondary holding period entirely by optimizing the conditions in the primary continuous bioreactor. However, further gains can be made using the holding period, by optimizing the holding temperature (diacetyl removal by yeast is very temperature dependent), the amount of fermentable sugars remaining in the liquid at the beginning of the holding period, optimizing the concentration of yeast present, the hydrodynamic characteristics of the holding vessel (diacetyl removal could be improved by improving the contact between the yeast and the beer), and taking further measures to eliminate oxygen from this stage.
The process described in this section, with a continuous bioreactor operating with a 24 hour residence time followed by a 48 hour batch hold, is 1.8 times more productive than a current industrial batch process. A relatively fast industrial batch process with a 7.5 day cycle time has a volumetric beer productivity of 0.093 m3 beer produced/(m3 vessel volume×day), whereas the continuous process described here has a productivity of 0.165 m3 beer produced/(m3 vessel volume×day). If further research allowed the batch holding period to be shorted to 24 hours, beer productivity would become 2.3 times more productive that the industrial batch standard. If the ideal scenario of a 24 hour continuous process with no batch holding were achieved, beer volumetric productivity would become 7.5 times that of the batch standard. In addition to the increased volumetric productivity, the additional benefits realized by moving from a batch to a continuous process, such as shorter time to market, decrease in brewhouse size, and less frequent yeast propagation, must be balanced with a careful analysis of relative operating costs.
Other researchers (Kronlöf and Virkajärvi, 1996; Nakanishi et al., 1993; Yamauchi et al., 1995) have focused on developing multi-staged continuous fermentations in which the first stage of continuous fermentation (aerobic) results in only a partial consumption of the fermentable sugars present in the wort. While this strategy has shown some success in terms of flavor production, these systems are complex. As well, the first aerobic stage of such systems creates an environment, which is more susceptible to microbial contamination (i.e. high sugar concentration, temperature, and oxygen, with low concentrations of ethanol). In the gas-lift bioreactor system presented in this work, the bioreactor has a low fermentable sugar concentration, low pH, high ethanol concentration, and low concentrations of oxygen, making the environment inhospitable for potential contaminants.
Effect of Liquid Residence Time on Key Yeast Metabolites During Continuous Primary Beer Fermentation
FIGS. 148 to 153 show the analytical results obtained from the bioreactor liquid phase. In Table 17, the average concentrations and flow rates of the measured analytes at pseudo-steady state (after a minimum of three bioreactor turnover times) are listed at the two liquid residence times used during this experiment. While liquid phase yeast viability did not change significantly when the flow rate of wort to the bioreactor was increased, the concentration of yeast cells did change as seen in
In
The liquid phase concentration of the fermentation product ethanol decreased with the step change in liquid residence time. However, the system as a whole was producing more ethanol on a mass flow rate basis at the faster liquid residence time. Because the objective of this invention was not only to produce ethanol in isolation, but rather a beer with a balance of many components, maximizing ethanol productivity must be balanced with other factors. At the end of a commercial primary beer fermentation, the majority of fermentable glucose substrate must be consumed.
In
Ethyl acetate and isoamyl acetate mass flow rates given in Table 17(b) both increased in response to the decrease in liquid residence time. In
Using a Commercial Preparation of Alpha-Acetolactate Decarboxylase to Reduce Total Diacetyl During Continuous Primary Beer Fermentation
Trial 1: As a result of numerous bioreactor upgrades, the system operated without contamination throughout the duration of this trial. The data for this trial are given in FIGS. 154 to 156. In Table 19 the average pseudo-steady state concentrations of total diacetyl before and after ALDC addition to the wort are summarized. Total diacetyl concentration dropped by 47 percent with the addition of ALDC to the wort, which makes the use of this enzyme promising for the future (averages taken after three bioreactor turnover times). As seen in
In order to eliminate the potential confounding effect of wort variability, during Experiment 3 a large quantity of wort from the brewhouse (14 hL) was collected and ALDC was added directly to the wort remaining in this holding vessel, once a pseudo-steady state baseline was reached. This further eliminated any potential wort inconsistencies that could have affected fermentation performance in Trial 1. This wort storage vessel was also equipped with carbon dioxide sparging, so that dissolved oxygen levels in the wort supply were kept at a consistently low level.
Trial 2: FIGS. 157 to 159 illustrate the effect of ALDC addition to the wort supply, on total diacetyl, total fermentable carbohydrate (as glucose), ethanol, and the freely suspended cell concentration during continuous beer fermentation. No contamination was detected at any point during this experiment. The concentration of total diacetyl was reduced by 45 percent upon addition of ALDC. No significant differences in ethanol, total fermentable carbohydrate (as glucose) or the freely suspended cell concentration were observed, which agrees with the batch findings of Aschengreen and Jepsen (1992).
The results of Trials 1 and 2 indicate that ALDC did have a significant effect on total diacetyl concentration during continuous fermentation in gas lift bioreactors giving an average reduction in total diacetyl concentration of 46 percent. This has the potential to decrease, or eliminate secondary processing for diacetyl reduction in continuous gas-lift systems. A relatively high dosage of ALDC was used for these initial experiments, and it would be necessary to optimize the amount, method and timing of ALDC dosing, in wort if this enzyme was to be adopted for the process. Further savings could be realized if an enzyme becomes available with higher activity levels under brewery fermentation conditions or if the enzyme itself was immobilized, thus allowing for its reuse (Dulieu et al., 1996). Another consideration will be public acceptance of enzyme additives that have been produced using genetically modified organisms.
At the supplier's recommended dosage of 2 kg/1000 hL, and, with cost of the commercial enzyme preparation at $131.05/kg, $0.26/hL would be added to the material costs of fermentation. As used in the experiments performed, the enzyme dosage was 72 μg/L (108 ADU/L) or 7.2 kg/1000 hL ALDC, giving an added material cost of $0.94/hL. The economics of using ALDC for diacetyl reduction during gas lift continuous fermentations will depend on the optimum enzyme dosage under bioreactor conditions and the amount of time saved by its use.
*averages based on pseudo-steady state values after three reactor turnover times
The foregoing supports the proposition that continuous fermentation using immobilized yeast and the associated free cells in a gas-lift draft tube bioreactor system is a viable alternative to batch fermentation for beer production based on the following criteria:
There are still many areas which need to be studied further, but the technology is ready to be tested at a larger scale. Gas-lift bioreactors are already used at an industrial scale for wastewater treatment, which makes the prospects of scaling up the continuous beer fermentation system technically feasible. The Grolsch brewery in the Netherlands has been reported to use a 230 m3 gas lift bioreactor for treatment of their wastewater (Driessen et al., 1997). One of the biggest barriers to commercial scale continuous fermentation in the brewing industry may be the acceptance by the brewers of a new process, in an industry that is deeply tied to tradition.
Data collected on secondary yeast metabolites produced during continuous beer fermentations conducted in this work highlighted the importance of controlling oxygen in the fluidizing gas for beer flavor formation. The findings showed that under the given operating conditions, increased air in the bioreactor fluidizing gas caused an increase in acetaldehyde, diacetyl, and higher alcohols (isoamyl alcohol and isobutanol), while the concentrations of esters (isoamyl acetate, ethyl hexanoate, ethyl octanoate) and ethanol were reduced. These data suggest that there is the potential for controlling beer flavor through the composition of the bioreactor fluidizing gas, allowing for the production of unique products.
With the exception of when air was eliminated from the fluidizing gas, a freely suspended cell concentration of greater than 108 cells/mL was maintained in the bioreactor liquid phase. The system thus has more than one population of yeast cells coexisting in the bioreactor, the immobilized yeast and the liquid phase suspended yeast. Because of the large quantities of viable yeast growing in the bioreactor liquid phase, the possibility exists of using a continuous bioreactor as a yeast propagator.
When a secondary 48-hour batch-holding period was added following continuous primary fermentation, a flavor profile within the range of market beers was obtained. The temperature of this holding period was 21° C., and the importance of minimizing the exposure of the liquid to oxygen during the holding period for flavor formation was demonstrated experimentally. The addition of a holding period adds two days to the process as well as additional complexity. However, it is still significantly faster than commercial batch fermentations, which take between seven and fourteen days.
Ultimately, the ideal scenario would be to entirely eliminate the secondary holding period by optimizing the conditions in the primary continuous bioreactor. However, further reductions in the secondary holding time can be achieved in the short term by optimizing the holding temperature (diacetyl removal by yeast is very temperature dependent), the amount of fermentable sugars remaining in the liquid at the beginning of the holding period, the concentration of yeast present, the hydrodynamic characteristics of the holding vessel (diacetyl removal could be improved by improving the contact between the yeast and the beer), and by taking further measures to eliminate oxygen from this stage.
Other researchers (Kronlof and Virkajarvi, 1996; Nakanishi et al., 1993; Yamauchi et al., 1995) have focused on developing multi-stage continuous fermentations in which the first stage of continuous fermentation (aerobic) results in only a partial consumption of the fermentable sugars present in the wort. While this strategy has shown some success in terms of flavor production, these systems are complex. As well, the first aerobic stage of such systems creates an environment, which is more susceptible to microbial contamination (i.e. high sugar concentration, temperature, and oxygen, with low concentrations of ethanol). In the gas-lift bioreactor presented in this work, the bioreactor has a low steady state fermentable sugar concentration, low pH, high ethanol concentration, and low concentrations of oxygen, making the environment inhospitable for potential contaminants. In a less-developed brewery, minimizing complexity and developing a robust, contamination-resistant process is an important success factor.
The addition of a commercial preparation of alpha-acetolactate decarboxylase (ALDC) to the wort supplying the continuous fermentation showed an average diacetyl reduction of 46 percent. However, because ALDC is an enzyme that is produced by a genetically modified organism (GMO), there are public perception issues that would need to be addressed before using such an enzyme in a commercial product. In addition the commercially available enzymes for diacetyl control do not have optimal activity under fermentation conditions.
Over six months of continuous fermentation using kappa-carrageenan gel immobilization, freely suspended cells in the liquid phase retained viabilities greater than 90 percent, while immobilized cell viability decreased to less than 60 percent. Scanning electron micrographs revealed that cells located near the periphery of the gel bead had multiple bud scars and a regular morphology, while those near the bead core had an irregular shape and fewer bud scars, suggesting impaired growth. The micrographs also suggested that the yeast located near the bead core were showing signs of aging. As discussed in section 5, kappa-carrageenan gel has many characteristics that make it a desirable yeast immobilization matrix. However, there is currently no industrial method available for bead manufacture and, because the yeast are entrapped in the matrix as part of the bead-making process, bead-handling in a commercial plant increases complexity and cost. Other immobilization methods such as self-aggregation or flocculation should be explored in the future. This would eliminate the complexity of bead handling in a plant environment, and, if the yeast flocs were disrupted on a regular basis, one could ensure that aged cells are regularly removed from the bioreactor.
This application is a continuation-in-part of co-pending application Ser. No. 10/175,351 filed Jun. 20, 2002.
| Number | Date | Country | |
|---|---|---|---|
| Parent | 10175351 | Jun 2002 | US |
| Child | 11098688 | Apr 2005 | US |
