United StatesEnvironmental ProtectionAgency

Air

Office of Air QualityPlanning and StandardsResearch Triangle Park NC 27711

EPA-450/3-91-027January 1992

Assessment of VOC Emissionsand Their Control from Baker’sYeast Manufacturing Facilities

ASSESSMENT OF VOC EMISSIONS AND THEIR CONTROLFROM BAKER'S YEAST MANUFACTURING FACILITIES

Control Technology Center

Sponsored by:

Emission Standards DivisionOff ice of Air Qual i ty Planning and Standards

U. S. Environmental Protection AgencyResearch Triangle Park, North Carolina 27711

Air and Energy Engineering Research LaboratoryOffice of Research and DevelopmentU. S. Environmental Protection Agency

Research Triangle Park, North Carolina 27711

January 1992

EPA No. 450/3-91-027January 1992

ASSESSMENT OF VOC EMISSIONS AND THEIR CONTROLFROM BAKER'S YEAST MANUFACTURING FACILITIES

BY

Midwest Research Institute401 Harrison Oaks Boulevard

Suite 350Cary, North Carolina 27513

EPA Contract Nos. 68-Dl-0115,68-DO-0137, and 68-02-4379

Project LeadMartha Smith

Off ice of Air Qual i ty Planning and StandardsU.S. Environmental Protection Agency

Research Triangle Park, North Carolina 27711

Prepared for:

Control Technology CenterU. S. Environmental Protection Agency

Research Triangle Park, North Carolina 27711

NOTICE

This report was prepared by Midwest Research Institute,Cary, North Carolina. It has been reviewed for technicalaccuracy by the Emission Standards Division of the Office of AirQuality Planning and Standards and the Air and Energy EngineeringResearch Laboratory of the Office of Research and Development,U. S. Environmental Protection Agency, and approved forpub l ica t ion. Mention of trade names or commercial products isnot intended to constitute endorsement or recommendation of use.

i i

ACKNOWLEDGEMENT

This report was prepared for EPA's Control Technology Center(CTC) by Robin Barker and Maresa Williamson of Midwest ResearchI n s t i t u t e . The Work Assignment Manager was Martha Smith of EPA'sChemicals and Petroleum Branch (CPB). Also par t i c i pa t i ng on theproject team were Bob Blasczcak, CTC; Chuck Darvin, EPA Air andEnergy Engineering Research Laboratory; and K.C. Hustvedt, CPB.

i i i

PREFACE

The Control Technology Center (CTC) was established by theU. S. Environmental Protect ion Agency's (EPA's) Off ice ofResearch and Development and Office of Air Quality Planning andStandards to help State and local air pol lut ion control agenciesimplement their air toxics and other pol lut ion programs. Threelevels of assistance can be accessed through the CTC. F i r s t , aCTC Hotline has been established to provide telephone assistanceon mat ters re la t ing to a i r po l lu t ion cont ro l technology. Second,more in-depth engineering assistance can be provided whenappropriate. Third, the CTC can provide technical guidancethrough publication of technical guidance documents, developmentof personal computer software, and presentation of workshops oncontrol technology matters.

The major objectives of this document are to provide ageneral overview of the baker's yeast production process, tosummarize available data on VOC emissions from baker's yeastmanu fac tu r i ng fac i l i t i es , and to evaluate potent ial emissioncontro l opt ions. This work was initiated by a State of Marylandrequest for technical support from the CTC. The State isconcerned that while VOC concentrations in the flue gas from theyeast fermentors a t a fac i l i ty in Ba l t imore, Mary land, are low,the large volume of flue gas indicates a high mass emission rate.Thus, the State of Maryland is looking for guidance oncon t ro l l i ng these d i l u te , high-volume gas streams.

i v

TABLE OF CONTENTS

1.0 INTRODUCTION............. ...............................................................................

1.1OVERVIEW...............................................................................................................

2.0 INDUSTRY PROFILE................................................................................................

2.1 EVOLUTION OF THE INDUSTRY .................................................................2.2 NUMBER AND LOCATION OF FACILITIES ..................................2.3 OZONE NONATTAINMENT STATUS .....................................................2.4 PRODUCTION STATISTICS AND GROWTH PROJECTION .......................

3.0 PROCESS DESCRIPTION..................................................................

3.1 RAW MATERIALS..............................................................3.2 FERMENTATION...............................................................

4.0

3.2.1 General ...........................................................................3 .2 .2 Fermentation Sequence3.2.33.2.4

Laboratory Stage, o r F l a s k S t a g e ( F 1 )Pure Culture Stages (F2 and F3) ......................

3.2.5 M a i n F e r m e n t a t i o n S t a g e s ( F 4 - F 7 ) . . . . . .3 . 2 . 6 Harvesting and Packaging ..............................................3 .2 .7 Production of Dry Yeast ......................................

EMISSION ESTIMATIONS ..............................................................

4 . 1 SOURCES OF EMISSIONS...............................................................4.2 EMISSION DATA AND PROCESS EMISSION FACTORS..................4 . 3 NATIONWIDE EMISSION LEVELS ..............................................4 . 4 MODEL EMISSION STREAM ....................................................

5.0 EMISSION CONTROL TECHNIQUES .....................................................

6.0

5 . 1 C O M P U T E R I Z E D P R O C E S S C O N T R O L M E A S U R E S5.2 ENGINEERING CONTROLS..............................................................

5.2.1 Wet Scrubbers (Gas Absorption)....................................5.2.2 Carbon Adsorption .........................................................5.2.3 Incineration............ ............................................................5.2.4 Condensation .....................................................................5.2.5 B io l og i ca l F i l t r a t i on

5.3 IMPACT ASSESSMENT OF CONTROL OPTIONS....................................

CONCLUSIONS ...................................................................................

7.0 REFERENCES .......................................................................................

V

Page

1

1

3

3355

5

710

10101212131517

17

18192323

24

2427

273030333537

41

44

Figure 1.

Figure 2.

Figure 3.

Figure 4.

Figure 5.

TABLE 1.

TABLE 2.

TABLE 3,

TABLE 4.

TABLE 5.

TABLE 6.

TABLE 7.

TABLE 8.

LIST OF FIGURES

Process flow diagram for producing baker'sy e a s t . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Fermentation sequence used in producingb a k e r ' s y e a s t . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

Emissions profile of a typical tradef e r m e n t a t i o n . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

S c h e m a t i c o f a p a c k e d t o w e r a b s o r b e r . . . . . . . . .

Schematic of a biofiltration system................

LIST OF TABLES

Y E A S T M A N U F A C T U R I N G P L A N T S . . . . . . . . . . . . . . . . . . .

OZONE NONATTAINMENT STATUS...................................

VOC EMISSIONS FROM YEAST FERMENTATION.........................

VOC EMISSIONS FROM A TYPICAL YEASTFERMENTATION FACILITY .. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

SUMMARY OF THE COST AND ENVIRONMENTALIMPACTS OF THE CONTROL OPTIONS ..........................

CAPITAL COSTS OF ADD-ON CONTROL OPTIONS...............................

ANNUAL COSTS‘OF ADD-ON CONTROL OPTIONS ....................

COSTS FOR CONDENSER CONTROL OPTION ............................

Page

8

11

25

28

36

4

6

20

22

38

39

40

42

v i

1.0 INTRODUCTIONThe object ives of this study were to obtain information on

baker's yeast fermentation processes, to determine the number andlocat ions of yeast p lants , to est imate the potent ial emissionsfrom the process, and to evaluate potential emission controlopt ions. The information contained in this report includes ani n d u s t r y p r o f i l e , descript ions of the production process,avai lable emission data, descript ions and technical evaluat ionsof the cont ro l opt ions, the impacts (cost and environmental) ofeach option, and the results of an evaluation to determine themost effect ive and least cost ly technology opt ions.

This document is organized into the fol lowing sect ions:Sect ion 2.0 prov ides a br ie f character izat ion of the baker 'syeast manufacturing industry, including the number, locations,and product ion capac i t ies o f fac i l i t ies ; Sect ion 3 .0 descr ibesthe baker's yeast fermentat ion process, including the equipmentused, feed materials and process requirements, and typicalprocess y ie lds and l imi ta t ions; Section 4.0 presents the emissionsources and pollutant types, avai lable emission test data,process emission factors, nationwide emission levels, and modelemission stream data; Sect ion 5.0 presents the avai lablein format ion on potent ia l emiss ion cont ro ls for the po l lu tants andemission points ident i f ied in Sect ion 4.0; Sect ion 6.0 summarizesthe conclusions derived-from this invest igat ion; and Section 7.0prov ides re ferences for th is repor t .1.1 OVERVIEW

Current ly 13 fac i l i t ies produce baker 's yeast in the Uni tedStates. The faci l i t ies are widely dispersed and can be found in10 d i f ferent States. Ten of the fac i l i t ies are located in ozonenonattainment areas. The majority of yeast manufacturingfacilit ies employ a moderate degree of process control to reducethe amount of volatile organic compounds (VOC's) generated.However, on ly one fac i l i ty has appl ied an a i r po l lu t ion cont ro lsystem to control emissions from the process. Because of thelocat ion o f some fac i l i t ies in nonat ta inment areas, cont ro l l ingyeast production VOC emissions may help bring these areas into

1

compliance with the national ambient air quality standard (NAAQS)for ozone.

Baker's yeast is produced by a fermentation process in whichlarge quantities of ethanol and acetaldehyde are generated andemitted to the atmosphere. Based on test data from threef a c i l i t i e s , the VOC mass emission rate from a typical faci l i ty isestimated at 82 megagrams per year (Mg/yr) (90 tons pery e a r [ t o n s / y r ] ) . Ethanol is approximately 80 to 90 percent ofthe emissions generated, and the remaining 10 to 20 percentconsists of other alcohols and acetaldehyde. The VOCconcentrat ion from a typical trade fermentor varies over a rangefrom 5 to 600 parts per mill ion by volume (ppmv). Thef luc tuat ion in the VOC concentra t ion is a t t r ibutab le pr imar i ly tothe var ia t ion in feed rates to the fermentor , var ia t ion in thea i r f l o w r a t e , and the design of the fermentor 's air sparger andagitat ion systems. Based on the number of yeast facilities andthe typical VOC emission levels, the total annual VOC emissionsfrom this source category are estimated to range from 780 to1,060 Mg/yr (860 to 1,170 tons/yr).

The VOC emission alternatives that were evaluated duringthis study were process control measures to reduce the formationof VOC emissions as well as wet scrubbers, carbon adsorbers,incinerators, condensers, and b io log ica l f i l te rs to cont ro l VOCemissions. Of these approaches, i t appears that process controlmeasures, c a t a l y t i c i n c i n e r a t o r s , or a combination of add-oncontro l techniques (e.g. , wet scrubbers followed by aninc ine ra to r o r a b io log i ca l f i l t e r ) a re t he mos t f eas ib leapproaches for control l ing yeast process emissions. Based on theresu l t s o f t h i s s tudy , the cont ro l e f f ic iency assoc ia ted wi th theadd-on control systems is estimated to be 95 to 98 percent.

Process control measures would limit the amount of ethanolformed by the process. This can be accomplished by incrementallyfeeding the molasses mixture (the principal source of carbon foryeast growth) and by supplying suff icient oxygen to thefermentors. Although these process control strategies arerout ine ly employed at yeast product ion fac i l i t ies for economic

2

reasons ( i .e . , to opt imize the use of raw materials), these stepsare not optimized to limit VOC emissions. Implementing andoptimizing more str ingent process control , especial ly during theearly stages of fermentation where close processusual ly not requi red, would reduce the formationby 75 to 95 percent.2.0 INDUSTRY PROFILE

c o n t r o l i sof VOC emissions

Information on the evolut ion of the baker's yeast indust ry ,the number and location of faci l i t ies, the ozone nonattainments t a t u s o f f a c i l i t y l o c a t i o n s , yeast product ion s ta t is t ics , andgrowth project ions for the industry is presented below.2.1 EVOLUTION OF THE INDUSTRY

The production of baker's yeast is a process that has beendeveloped over centuries. The earliest means of producingbaker's yeast was to inoculate fresh dough with a port ion of thepreceding fermented dough. Yeas t w i l l g row i nde f i n i t e l y by t h i smeans, and this method is st i l l used today in producing sourdough breads. Before the 1800's, the top-fermenting yeast frombreweries was used for baking bread. However, the yeast y ie ldswere very low and the ethanol yield was very high. Over the last100 years, grain mashes were replaced with molasses as theprincipal carbon source for producing yeasts, fermentat ionprocess tanks were equipped with air supply and incremental feedsystems to reduce the formation of ethanol and increase thequanti t ies of yeast produced. A t th i s po in t , t he yeas tproduction process became independent of the brewery process andhas remained virtually unchanged since the 1920's.2.2 NUMBER AND LOCATION OF FACILITIES

Six major companies manufacture baker's yeast in the UnitedStates. These companies are Universal Foods (Red Star Yeast),Fleischmanns, Gist-brocades, Lallemand (American Yeast), Minn-dak, and Columbia. There are a total of 13 manufacturing plantsowned by these companies in the United States. Tab le 1 l i s ts thelocations of the plants by manufacturer.

TABLE 1. YEAST MANUFACTURING PLANTS

Lallemand (American Yeast) Baltimore, MarylandColumbia Headland, AlabamaFleischmanns Gastonia, North Carol ina

Memphis, TennesseeOakland, Cal i fo rn iaSumner, Washington

Gist-brocades Bake rs f i e l d , Ca l i f o rn iaEast Brunswick, New Jersey

Minn-dak Wahpeton, North DakotaUniversal Foods Corp. (Red Baltimore, MarylandStar Yeast) Dallas, Texas

Milwaukee, WisconsinOakland, Ca l i f o rn ia

4

2.3 OZONE NONATTAINMENT STATUSThe baker's yeast manufacturing plants in the United States

are located in both attainment and nonattainment areas for ozone.This rank ing is accord ing to the la test update of the c i t iesexceeding the ozone limit, made in 1988.¹ The ozone air qual i tystandard calls for no more than 1 hour in a year during which anymonitor in an area records an O3 level in excess of 0.12 ppm.¹Table 2 presents the locations of al l the yeast manufacturingplants located in the United States and whether or not they arelocated in ozone nonattainment areas.2.4 PRODUCTION STATISTICS AND GROWTH PROJECTION

In 1989, only 12 yeast plants were in operat ion. The tota lU.S. production of baker's yeast in 1989 was 223,500 Mg(245 ,000 tons ) . O f th i s to ta l , approximately 85 percent of theyeast was compressed or cream yeast, and the remaining 15 percentwas dry yeast.

Between 1990 and 1991, two add i t iona l fac i l i t ies were openedand are currently in production, and one faci l i ty was closeddown. A foreign manufacturer of baker's yeast, Minn-dak, hasrecently opened a new yeast plant in Wahpeton, North Dakota.Fleischmanns also has opened a plant in Memphis, Tennessee, butc losed i ts p lant in St . Lou is , Missour i . The opening of thesetwo fac i l i t ies in 1990 is expected to increase product ion o fbaker's yeast by approximately 5 to 10 percent.3.0 PROCESS DESCRIPTION

Two main types of baker's yeast are produced: compressedyeast and active dry yeast (ADY). Compressed yeast is aperishable commodity and must be refrigerated or frozen at allt imes. Refrigerated compressed yeast remains useful for severalweeks before molds begin to develop. Frozen compressed yeast canbe stored and used for up to a month, but some softening of theyeast cake occurs. Act ive dry yeast has a lower bake act iv i tythan compressed yeast, but i t can be stored for 1 to 2 yearsw i thou t re f r i ge ra t i on be fo re the bake ac t i v i t y i s l os t .Compressed yeast is sold mainly to wholesale bakeries, whereasADY is sold mainly for home baking needs and, to a limited

5

TABLE

P l a n t l o c a t i o n

M i l w a u k e e , W I

O a k l a n d , C A

B a l t i m o r e , M D

D a l l a s , T XB a k e r s f i e l d , C AEast Brunswick, NJ

S u m n e r , W AM e m p h i s , T N

Gastonia, NCHeadland, AL

Wahneton. ND

2. OZONE NONATTAINMENT STATUSII Nonattainment

Company for ozoneUniversal Foods Corporation Yes(UFC)UFC YesFleischmanns YesLallemand (American Yeast) YesUFC YesUFC YesSist-brocades YesGist-brocades YesFleischmanns NoFleischmanns YesFleischmanns Yesa

Columbia NoYinn-dak No

aGastonia is not formally recognized by EPA as an ozonenonattainment area, but exceedances for ozone occur, and thearea will most likely be nonattainment for ozone when the 1990Clean Air Act Amendments are fully implemented.

6

extent , to bakeries. Compressed yeast and ADY are produced in asimilar manner, but ADY is dried after processing and isdeveloped from a dif ferent yeast strain. Another dry yeastproduct, instant dry yeast (IDY), is dr ied after processing andis produced from a faster-reacting yeast strain than that usedfor ADY. The main difference between ADY and IDY is that IDYdoes not have to be dissolved in warm water prior to usage,whereas ADY does. The fol lowing discussion is directed towardscompressed yeast manufacturing, a l though a br ie f d iscuss ion ofthe production of dry yeast is also presented (Section 3.2.7).

A variety of processes are used in producing baker's yeast.Most processes, however, are a variat ion on the Zulauf process,which was introduced in the early 1900's. This report provides ageneral descript ion of the Zulauf production process.

Figure X presents a process flow diagram for the productionof baker's yeast. The f i rs t s tages of product ion cons is t o fgrowing the yeast from the pure yeast culture in a series offermentat ion vessels. The yeast is then recovered from the f inalfermentor using centr i fugal act ion to concentrate the yeastso l i ds . The yeast product is subjected to one or more washingsin the cent r i fugal separator . The yeast so l ids are then f i l te redby a f i l t e r p ress o r a ro ta ry vacuum f i l t e r t o fu r the rconcentrate the yeast. Next, the yeast f i l te r cake is b lended inmixers with small amounts of water, emulsi f iers, and cutt ingo i l s . Af ter mix ing, the mixed press cake is extruded and cut.The yeast cakes are then either wrapped for shipment or dried toform dry yeast.3.1 RAW MATERIALS

The principal raw materials used in producing baker 's yeastare the pure yeast culture and molasses. The yeast strain usedin producing compressed yeast is Saccharomyces cerevisiae. 2 Caneand beet molasses are used as the principal carbon source topromote yeast growth. Molasses contains 45 to 55 percentfermentable sugars by weight in the forms of sucrose, glucose,and fructose. This sugar by-product is the least expensivesource of sugar known. Other sources, such as corn gri ts,

7

Raw Materials

Ethanol, CO2Fermentation 4

Filtration

Extrusion & Cutting

Packaging

Shipment of packaged yeast

Figure 1. Process flow diagram for producing baker's yeast.

8

r a i s i n s , or sugar-containing wastes of the confect ionary industryare a lso e f fec t ive , but for various reasons (mostly economic),these alternatives were found to be unsuitable as carbon andenergy substrates for baker's yeast production. 3

The amount and type of cane and beet molasses used dependson the avai labi l i ty of the molasses types, cost, and the presenceo f i nh ib i to rs and tox ins . Usual ly, a blend consist ing of bothcane and beet molasses is used in the fermentations. Onceblended, the molasses mixture is clarified to remove any sludge.P r i o r t o c l a r i f i c a t i o n , the pH of the molasses mixture isadjusted because too high a pH will promote bacterial growth.Bacterial growth occurs under the same conditions as yeastgrowth, making pH monitoring a very important factor. Thec lar i f ied molasses mix ture is then s ter i l i zed wi th h igh-pressuresteam. A f t e r s t e r i l i z a t i o n , i t i s d i lu ted wi th water and he ld inholding tanks unti l i t is needed by the fermentat ion process.

Other required raw materials are a variety of essentialnutr ients and vitamins. Mineral requirements include nitrogen,potassium, phosphate, magnesium, and calcium. Ni t rogen isnormally supplied through the addition of ammonium salts, aqueousammonia, or anhydrous ammonia to the feed stock.4 The molassesnormally provides suff ic ient quanti t ies of potassium and calcium.Phosphates and magnesium are added in the form of phosphoric acidor phosphate salts and magnesium salts.5 I ron, z inc, copper ,manganese, and molybdenum are also required in trace amounts.

Several vi tamins are required for yeast growth (biot in,inos i to l , pantothenic ac id , and th iamine) . Yeast wi l l not growin the absence of b io t in .6 Thiamine is not required for yeastgrowth but is normally added to the feed stock because it is apotent st imulant for fermenting doughs. Both cane and beetmolasses usual ly provide enough inositol and pantothenic acid foryeast growth. However, if beet molasses, wh ich i s de f i c i en t i nb i o t i n , is used, biotin must be added or a mixture of cane andbeet molasses is required.

3.2 FERMENTATIONYeast cel ls are grown in a series of fermentat ion vessels.

A typical fermentat ion process is shown in Figure 2. The processbegins when a pure yeast culture is grown in the laboratory.Por t ions of th is pure cu l ture are p laced in the f i rs t fermentoralong with the other feed materials and are allowed to grow.Yeast is propagated when the entire yeast mixture (or a port ionof the mixture from the preceding fermentor) is placed into thenext fermentor, which is equipped for batch or incrementalfeeding of the molasses malt. The process continues unti l theyeast mixture reaches the f inal fermentat ion vessel.3 .2.1 Genera l

Yeast fermentation vessels are operated under aerobiccondit ions (free oxygen present, or excess air) because underanaerobic condit ions ( l imited or no oxygen avai lable) thefermentable sugars are consumed in the formation of ethanol andcarbon dioxide, which resu l ts in low yeast y ie lds .7 Yeast yieldsunder anaerobic condit ions are often less than 10 percent-by-weight of fermentable sugars, whereas yeast yields of up to50 percent-by-weight of fermentable sugars are obtained underaerobic condit ions. 7 Therefore, to max imize yeast y ie lds , i t i simportant to supply enough oxygen for the dissolved oxygencontent in the l iqu id surrounding the yeast ce l ls to be at anopt imal leve l . I n p rac t i ce , oxygen transfer rates are ofteninadequate, and under such conditions, some ethanol is formed.In add i t i on , i t is a lso impor tant to cont ro l the amount o ffermentable sugars present in the fermentor, so that the sugar isassimi lated by the yeast as fast as i t is added. This balance isaccomplished by using an incremental feed system in the finalfermentat ion stages.3.2.2 Fermentation Sequence

Compared with subsequent fermentat ion stages, in the f irststages of yeast propagation, the medium is r icher in nut r ients ,and there is less aerat ion. Consequently, the fermentor l iquorcontains more alcohol and yields of yeast are lower. The loweryields in the f i rst stages are not necessari ly a drawback,

10

because the overall economy of the operation depends on the yieldf rom the f ina l t rade fermentat ion s tage. The fol lowing sect ionsdescribe each stage in the fermentation sequence.3.2.3 Laboratory Stage, or Flask Stage (F1)

The f i rs t fermentat ion s tage takes p lace in the laboratorywhen a port ion of the pure yeast culture is mixed with themolasses malt in a ster i l ized Erlenmeyer or Pasteur f lask. Theto ta l con ten ts o f t he f l ask a re t yp i ca l l y l ess than 5 l i t e r s (L )(1 .3 ga l l ons [ga l ] ) , and the yeast is allowed to grow in thef lask for 2 to 4 days.8 The entire f lask contents are then usedto inoculate the second fermentat ion stage.3.2.4 Pure Culture Stases (F2 and F3)

General ly, th is s tage consis ts o f two pure cu l turefermentat ions. The capacit ies of the fermentat ion vessels usedin this stage range from 1,140 L (300 gal) to 26,500 L(7,000 gal) . The pure culture fermentations are batchfermentations where the yeast is allowed to grow for 13 to24 hours. The contents of the fermentor from the f irst pureculture stage (F2) is added to the next fermentat ion vessel,which already contains the nutr ient-r ich molasses malt. The purecul ture fermentat ions are bas ica l ly a cont inuat ion of the f laskfermentat ion, except that the pure culture fermentations haveprov is ions for s ter i le aerat ion and asept ic t ransfer to the nextstage. The yeast yield in the pure culture fermentat ions isapprox imate ly 27 k i lograms (kg) (60 pounds [ lb ] ) in the f i rs tfermentor and 600 kg (1,300 lb) in the second fermentor.

The c r i t i ca l f ac to r i n the pu re cu l tu re ope ra t i on i ss t e r i l i t y . Rigorous ster i l izat ion of the fermentat ion mediumprior to inoculation is conducted by heating the medium underpressure or by boi l ing i t at atmospheric pressure for extendedper iods. I f a s ter i le env i ronment is not prov ided, contaminat ingmicroorganisms can easily outgrow the yeast.

The need for process control in the pure culture medium isl im i t ed . However, microbiological test ing of the medium before,dur ing, and after each fermentat ion is essential . The maltconcentrat ion of the in i t ia l fermentor is o f ten s tandard ized to

12

between 5 and 7.5 percent sugar.9 Once the pure culturefermentat ion is s tar ted, the only control lable parameters are thetemperature and the degree of aeration. However, con t ro l o faerat ion is not cr i t ical because of the excess sugar present, andcontro l o f temperature is not cr i t ica l because the fermentat ionoperates over a broad temperature range.3.2.5 Main Fermentation Stases (F4-F7)

The majori ty of the yeast yield grows in the f inalfermentation stages. The main fermentation stage takes place intwo to four fermentat ion vessels. These yeast fermentors varyconsiderably in size. The volumes of fermentation vessels in theF4 through F7 stages range from 37,900 L (10,000 gal) to over283,900 L (75,000 gal). The vessels have diameters in excess of7.0 meters (m) (24.5 feet [ f t ]) and heights up to 14 m (45 ft) .The larger vessels are associated with the f inal fermentat ionstages (F6 and F7). The fermentat ion vessels are typical lyoperated at a temperature of 30°C (86°F).

The fermentors are usual ly constructed of stainless steeland are equipped with an incremental feed system. Thisincremental feed system may be a pipe or a series of pipes thatdistr ibute the molasses over the entire surface of the fermentorl i q u i d . The rate at which the molasses is fed is cr i t ical andmay be controlled by a speed controller connected to a pump or bya valve on a rotameter, ‘which del ivers a certain volume ofmolasses at regulated t ime intervals. Nu t r i en t so lu t i ons o fvi tamins are kept in smal l , separate tanks and are chargedthrough rotameters into the fermentor, but the rate of feed isnot as cr i t ical as with molasses. However, if ammonia is used asa nitrogen source, additions must be made in a manner that avoidssudden pH changes. The nitrogen salts and phosphates may becharged in a shorter period of time than the molasses feed.Fermentors must also be equipped with heat exchangers to removethe heat produced from the production process and to cool thefermentor. The type of heat exchanger system used depends on thesize of the fermentat ion vessel. Because large volumes of airare suppl ied to the fermentat ion vessels during this stage of

13

production, the fermentor s ize in f luences the type of aerat ionsystem selected. The dif ferent types of aerat ion systems includehor izonta l , per forated p ipes; compressed air and mechanicala g i t a t i o n ; and a self-pr iming aerator.

In the hor izonta l , per forated p ipe system, a i r is b lownthrough a large number of horizontal pipes that are placed nearthe bottom of the fermentor. With this aerat ion system, the onlyagi ta t ion of the fermentor l iqu id is car r ied out by the act ion ofthe air bubbles as they r ise to the surface. T y p i c a l l y , t h i stype of aeration system requires from 25 to 30 m3 (880 to1,060 f t3 ) o f a i r to produce 0.45 kg (1 lb) o f yeast . 10

The eff ic iency of aerat ion with a given volume of air can begreat ly increased by mechanical agitat ion. In a compressedair/mechanical agitat ion aerat ion system, air under pressure issuppl ied to a c i rcu lar d i f fuser p ipe, Di rec t ly above the a i ro u t l e t s , a hor izonta l turb ine d isc prov ides mechanica l ag i ta t ion,which d is t r ibutes the a i r bubbles un i formly . Agitat ion systemshave baf f les to keep the fermentor l iqu id f rom rotat ing in thedi rect ion of the mot ion of the d isc . Th i s un i fo rm d i s t r i bu t i onof air bubbles reduces the volume of air needed to grow theyeast. In an agitated system, only 10 to 15 m3 (350 to 530 ft3)o f a i r are requi red to produce 0.45 kg (1 lb) o f yeast . 10

The self-pr iming aerator operates with a turbine that drawsair through a hol low, ve r t i ca l sha f t i n to the fe rmen to r l i qu id .Because air is drawn through the shaft of the turbine without acompressor, the pressure of the a i r a t the out le ts is not veryhigh and the depth to which the turbine can be submerged isl im i t ed .

When using the four-stage fermentat ion series, the pureculture stage is fol lowed by an intermediate stage (F4) of yeastgrowth without incremental feeding. The entire fermentorcontents from the intermediate stage are then pumped into a tankthat is equipped for incremental feeding and that has goodaerat ion. Th is s tage (F5) is o f ten ca l led s tock fermentat ionbecause, after fermentat ion is completed, the yeast is separatedf rom the bu lk o f the fermentor l iqu id by centr i fug ing, producing

14

a stock (p i tch) o f yeast for the next s tage. The third stage(F6) is usual ly carr ied out in fermentors as large as those usedfor the t rade fermentat ion or f ina l fermentat ion. Ae ra t i on i svigorous, and molasses and other nutrients are fed incrementally.The fermentor l iquor from this fermentor (F6) is usual ly dividedin to severa l par ts for p i tch ing the f ina l t rade fermentat ion(adding the yeast to s tar t fermentat ion) . A l ternate ly , the yeastmay again be separated-by centrifuging and stored for severaldays pr ior to i ts use in the f ina l t rade fermentat ions. Thefinal trade fermentat ion (F7) has the highest degree of aerat ion,and molasses and other nutrients are fed incrementally. Due tothe large a i r suppl ies requi red dur ing the f ina l t radefermentat ions, these vessels are often started in a staggeredfashion to reduce the load or size of the air compressors.requi red. The durat ion of each of the f inal fermentat ion stagesranges from 11 to 15 hours. The amount of yeast growth increasesin each s tage and is typ ica l ly 120 kg (265 lb) in the f i rs tstage, 420 kg (930 lb) in the second stage, 2,500 kg (5,510 lb)in the th i rd s tage, and 15,000 to 100,000 kg (33,070 to220 ,460 lb ) i n the fou r th s tage . 11

When using the two-stage f inal fermentat ion series, the onlyfermentat ions are the stock fermentat ion and the tradefermentat ion (F5 and F7, respect ively). About ha l f o f the13 yeast manufacturing -faci l i t ies use the four-stage f inalfermentat ion series and the other half use the two-stage process.

After al l of the required molasses has been fed into thefermentor , the l iqu id is aerated for an addi t iona l 0 .5 to1.5 hours. This permits further maturing of the yeast andresu l ts in a yeast that is more s tab le in re f r igerated s torage.3.2.6 Harvesting and Packaging

Once an optimum quantity of yeast has been grown, the yeastcel ls are recovered from the f inal trade fermentor by centr i fugalyeast separators. The separators used in this process arecont inuous dewater ing centr i fuges. 12

Af te r t he f i r s t passthrough the separators, a yeast sol ids content of 8 to 10 percentcan be acquired from a fermentor l iquor containing 3.5 to

15

4 . 5 p e r c e n t s o l i d s . 12

Next, the yeast is washed with water andpassed through the separators a second time. The second passusual ly produces concentrat ions of 18 to 21 percent so l ids. 12

I f the concentrat ion of yeast sol ids recovered from the secondpass is below 18 percent, then another washing and a third passthrough the separators is normally required. The yeast creamresult ing from this process can be stored for several weeks at atemperature slightly above 0°C (32°F). Af ter s torage, the yeastcream can be used to propagate yeast in other trade fermentationsor can be fur ther dewatered by f i l t ra t ion.

The centr i fuged yeast sol ids are further concentrated byp r e s s i n g o r f i l t r a t i o n . Two types of f i l ter ing systems are used:f i l te r presses and ro tary vacuum f i l ters . I n t h e f i l t e r p r e s s ,the f i l ter c lo th cons is ts o f cot ton duck or a combinat ion ofcot ton duck and synthet ic f ibers so t ight ly woven that no f i l teraid is necessary. Fi l ter presses having frames of 58 to115 centimeters (cm) (24 to 48 inches [in.]) are commonly used,and pressures between 860 to 1,030 kiloPascals (125 to 150 poundsper square inch) are appl ied. Yeast yields between 27 and12

32 percent solids may be obtained by pressing. 12

Rotary vacuumfi l ters are used for continuous feed of yeast cream. Generally,the f i l ter drum is coated with yeast by rotat ing the drum in atrough of yeast cream or by spraying the yeast cream directlyonto the drum. The f i l te r sur face is coated wi th potato s tarchcontaining some added salt to aid in drying the yeast product.The f i l ter drum rotates at a rate of 15 to 22 revolut ions perm inu te ( rpm) . 12

As the drum rotates, blades at the bottom of thedrum remove the yeast. Af ter a f i l te r cake of yeast is formedand while the drum continues to rotate, excess salt is removed byspraying a small amount of water onto the f i l ter cake. From thisprocess, f i l ter cakes containing approximately 33 percent sol idsare formed.

The f i l ter cake is blended in r ibbon mixers with smal lamounts o f water , emuls i f iers , and cut t ing o i ls . Emuls i f ie rs areadded to give the yeast a white, creamy appearance and to inhibitwater-spott ing of the yeast cakes. A small amount of oil,

16

usual ly soybean or cottonseed oi l , is added to help extrude theyeast. The mixed press cake is then extruded through open-throated nozzles to form continuous ribbons of yeast cake. Theribbons are cut and the yeast cakes are wrapped with wax paper.The wrapped cakes are cooled to below 8°C (46°F), at which timethey are ready for shipment in refr igerated trucks.3.2.7 Production of Dry Yeast

In yeast manufacturing, two types of dry yeast are produced:(1) ADY and (2) IDY. Active dry yeast is produced from a yeasts t ra in ident i f ied as No. 7752 in the American Type CultureC o l l e c t i o n . 13 This s t ra in g ives bet ter y ie lds than that used inproducing compressed yeast and is commonly referred to as the"d ry yeas t s t ra in . " Instant dry yeast is produced from a yeaststrain dif ferent from that used for ADY. The ADY and IDY areproduced through the same process as that described forcompressed yeast. A f t e r f i l t r a t i o n , the dry yeast product is.sent to an extruder, where emuls i f ie rs and o i ls d i f ferent f romthose used for compressed yeast are added to texturize the yeastand aid in extruding the yeast. Af ter the yeast is ext ruded in.thin r ibbons and cut, the yeast is dr ied in e i ther a batch or acontinuous drying system. Fluidized bed dryers can be used todry the extruded yeast. The extruded yeast strands are fed intothe drying chamber of a fluidized bed dryer. Heated air blowninto the bottom of the ‘dryer suspends the yeast part ic les into afluid bed and dries them. The drying t ime varies from 0.5 to4 h o u r s ( h r ) . 14 The humidity in the dryers is continuouslymonitored to determine when the drying cycle is complete.Fol lowing drying, the yeast is vacuum-packed or packed undernitrogen gas before heated sealing. The shelf life of ADY andIDY is 1 to 2 years at ambient temperature.

15

4.0 EMISSION ESTIMATIONSThe fol lowing sect ions present a composite of the avai lable

emissions data, process emission factors, nationwide emissionsestimates, and the documentation on the development of a modelprocess emission stream.

17

4.1 SOURCES OF EMISSIONSThe VOC emissions are generated as by-products of the

fermentat ion process. The two major by-products are ethanol,which is formed from acetaldehyde, and carbon dioxide. Other by-products consist of other alcohols and organic acids such asbutanol, isopropyl alcohol, 2,3 butanediol , and acetate. Theseby-products form as a result of excess sugar present in thefermentor or an insuff icient oxygen supply to the fermentor.Under these conditions, anaerobic fermentat ion occurs and resultsin the excess sugar being broken down to form alcohols and carbondiox ide. When anaerobic fermentation occurs, 2 moles of ethanoland 2 moles of carbon dioxide are formed from 1 mole of glucose.

Under anaerobic conditions, the ethanol y ie ld is increasedand yeast yields are decreased. In produc ing baker 's yeast i t isessent ia l to suppress ethanol format ion in the f ina l fermentat ionstages by incremental feeding of the molasses mixture and bysupplying suff icient oxygen to the fermentor.

The rate o f e thanol format ion is h igher in the ear l iers tages (pure cu l ture s tages) than in the f ina l s tages of thefermentation process. The earl ier fermentat ion stages are batchfermentors, where excess sugars are present and less aeration isused during the fermentation process. These fermentations arenot cont ro l led to the degree that the f ina l fermentat ions arecontrol led, because the majori ty of yeast growth occurs in thef ina l fermentat ion s tages and, therefore, there is no economicalreason for equipping the earl ier fermentat ion stages with thenecessary process control equipment.

Another potential emission source is the wastewatertreatment system used to treat process wastewaters. I f t h efaci l i ty does not employ an anaerobic biological t reatmentsystem, signif icant quanti t ies of VOC could be emitted from thisstage of the process. For more information on wastewatertreatment systems as an emission source of VOC's, please refer toan earlier CTC document on industrial wastewater treatmentsystems ent i t led "Industr ial Wastewater Volat i le Organic

18

Emissions-- Background Information for BACT/LAERDete rm ina t ions . " 16

4 . 2 EMISSION DATA AND PROCESS EMISSION FACTORSEmission test data were received from three yeast

m a n u f a c t u r i n g f a c i l i t i e s . 1 7 - 2 1 From a combination of these threef a c i l i t i e s , emission test data were avai lable for the last fourfermentation stages (F4-F7).

During the fermentation process, ethanol and acetaldehydeare not formed at constant rates; therefore, over the course ofthe fermentation, the concentrations of these compounds varysignificantly depending upon the amount of excess sugars presentand the combined effectiveness of the aeration and agitationsystems to supply suff icient oxygen throughout the fermentorvolume. A review of the emission test data showed that the VOCconcentrat ions d id vary s ign i f icant ly for each fermentat ion s tageand between different fermentors at a given stage. Thisvariation in emissions was expected between facilit ies because ofthe dif ferences in the feed systems and the size of thefermentors. However, even wi th in a g iven fac i l i ty , the emiss iondata vary from fermentor to fermentor because of the differencesin the design of the air sparger system and the placement ofbaff les and mechanical agitators within the fermentors.

Table 3 presents the VOC emission levels measured duringbatch cycles for each type of fermentat ion. The emission testdata were converted from total VOC concentrations to ethanolconcentrations since ethanol is the primary VOC compound emitted.Ethanol is approximately 80 to 90 percent of the emissionsgenerated, and the remaining 10 to 20 percent consists of otheralcohols and acetaldehyde.

A review of the emission data in Table 3 reveals thesignif icant var iat ion in VOC emission levels. To obtain moremeaningful data, the process emission factors presented inTable 3 were developed in an effort to normalize the f luctuat ionsin emission data between faci l i t ies and fermentors. However,there was s t i l l a s ign i f icant var ia t ion in the process emiss ionfactors . Upon reviewing process information obtained from these

19

TABLE 3. VOC EMISSIONS FROM YEAST FERMENTATION17-21a

F4 (Intermediate)

Fermentation stages

F5/F6 (Stock/pitch) F7 (Trade)

Concentration range, ppmv

Average concentration, ppmv

Maximum concentration, ppmv

Batch emissions, kg (lb)

Process emission factors

900-4,600 2-1,350 5-600

1,900-2,400 50-700 200

3,000-4,600 200-l,350 600

24-71 (53-156) 6-821 (13.1,810) 4.5-154 (10-340)

VOC’s emitted volume offermentor operating capacity, kg/L(lb/gal)

0.0011-0.0014 0.0001-0.003 0.000036-0.0006(0.009-0.012) (0.0008-0.025) (0.0003-0.005)

VOC’s emitted per amount of yeastproduced, kg/l,000 kg (lb/l,000 lb)

12-49 (12-49) 0.5-40 (0.5-40) 0.19-4.7 (0.194.7)

aTotal VOC emissions as ethanol.

20

f a c i l i t i e s , it was concluded that the low end of the data rangeis a t t r ibutab le to fac i l i t ies that have implemented a greaterdegree of process control or have improved fermentor designs overthose fac i l i t ies that represent the h igh end of the data range.Therefore, the typical emission levels and process emissionfactors presented in Table 4 were developed to represent atypical faci l i ty with a moderate degree of process control .Based on process control information obtained from yeastf a c i l i t i e s , i t i s be l ieved that the major i ty o f yeastmanufactur ing fac i l i t ies fa l l w i th in the emiss ion rangespresented in Table 4.

The typical emission levels for each fermentat ion stagereveal the process control changes between fermentation stages.The intermediate stage (F4) that fol lows pure culturefermentat ion& is either batch or fed-batch. The degree ofprocess cont ro l is not as s t r ingent for th is type of fermentat ionas i t is for trade fermentat ion because the yeast productionoutput f rom th is fermentor is not as cr i t ica l as that f rom thef ina l t rade fermentat ion. As a result , the emiss ion leve ls forthe intermediate fermentation are much higher than those for thetrade fermentat ion. However, total batch emissions from theintermediate fermentation stage are lower than those from thetrade fermentation stage due to the smaller fermentors used andthe lower production rate. The f inal three fermentat ions (F5-F7)are typical ly carr ied out in the same fermentors. The t ighterprocess control measures used during these fermentations resultin the lower emission levels.

The annual VOC emission rates presented in Table 4 weredeveloped based on the batch emissions from each fermentor andthe typical number of batches produced per year. As shown inTable 4, the major i ty of emissions are associated with tradefermentation as a result of the number of batches produced peryear. Trade fermentations account for 80 to 90 percent of theemissions generated from a given faci l i ty.

21

TABLE 4. VOC EMISSIONS FOR A TYPICAL YEASTMANUFACTURING FACILITY a

Typical ethanol emission levels

Fermentation stages

F4 (Intermediate) F5/F6 (Stock/pitch) F7 (Trade)

Concentration range, ppmv

Average concentration, ppmv

Maximum concentration, ppmv

Batch emissions, kg (lb)

Typical process emission factors

900-4,600 2-400 6-600

1,900 200 250

3,000-4,600 200-400 500-600

36.3 (80) 49.9 (110) 63.5 (140)

VOC’s emitted per volume of fermentoroperating capacity, kg/L (lb/gal)

0.0012 (0.010) 0.0004 (0.0035) 0.0005 (0.0045)

VOC’s emitted per amount of yeastproduced, kg/l,000 kg (lb/l,000 lb)

Operating time

15 (15) 5.0 (5.0) 3.5 (3.5)

No. of batches per week 3 4 20

No. of weeks per year

Annual VOC emissions rate, Mg/yr(tons/yr)

52 52 52

5.6 (6.2) 8.5 (9.4) 66 (73)

aTotal VOC’s as ethanol.

22

4.3 NATIONWIDE EMISSION LEVELSBased on the process emission factor of 0.0005 kilogram of

VOC's emitted per liter per batch (0.0045 pound per gallon perbatch) o f fermentor operat ing capac i ty , the typ ica l s ize(117,260 L {31,000 gal}) of trade fermentors, the number ofbatches processed per year (1,040) at each faci l i ty, and thenumber of yeast manufacturing faci l i t ies (13), the nationwide VOCemissions from manufacturing baker's yeast is estimated at 860Mg/yr (950 tons/yr).

Based on the process emission factor of 0.0035 kilogram ofVOC's emitted per kilogram (0.0035 pound per pound) of yeastproduced and the current nationwide annual production of baker'syeast (220 mil l ion ki lograms [490 mil l ion pounds]), thenationwide VOC emissions from yeast manufacturing is estimated at780 Mg/yr (860 tons/yr).4.4 MODEL EMISSION STREAM

A model emission stream was developed in order to evaluatecontrol device performance at yeast manufacturing faci l i t ies.The model emission stream was developed based on the combinedemission levels from f ive 117,260-L (31,000-gal) t radefermentors. Trade fermentation was selected because the majorityof emissions are generated from this stage of the process. Inadd i t i on , the major i ty o f the f ina l fermentat ions (F5-F7) arecarried out in the same fermentors. Therefore, any contro ltechnique appl ied to the trade fermentors would also controlemissions from the stock and pitching fermentation stages. Five117,260-L (31,000-gal) trade fermentors were selected becausethis was the average number and capacity of trade fermentors ata l l yeas t manu fac tu r i ng fac i l i t i es . Each trade fermentor wasassumed to have an air flow rate of 159 m /min (5,600 ft3/min),3

which was based on air flow rate data for actual tradefermentors. Therefore, the model emission stream has an air flowrate of 790 m3/min (28,000 ft3/min) and an average VOCconcentration of 200 to 300 ppmv. The average VOC concentrationswere determined based on the emission levels from actual trade

23

fermentors. An emiss ion prof i le for th is fermentat ion s tage ispresented in Figure 3.5.0 EMISSION CONTROL TECHNIQUES

Only one yeast manufacturing facility employs an add-onpollution control system to reduce VOC emissions from thefermentat ion process. The po l lu t ion cont ro l system at th isfac i l i ty cons is ts o f a wet scrubber fo l lowed by a b io log ica lf i l t e r . However, al l of the yeast manufacturers suppress ethanolformation through varying degrees of process control.

The process control measures consist of incrementallyfeeding the molasses mixture to the fermentors so that excesssugars are not present and of supplying suff icient oxygen to thefermentors to opt imize the dissolved oxygen content of the l iquidin the fermentat ion vessel. The fol lowing sect ions provide amore detailed discussion of the process control measures toreduce ethanol formation and provide information on thefeasibi l i ty of implementing add-on control devices to reduce orel iminate ethanol emissions from yeast fermentat ion vessels.5.1 COMPUTERIZED PROCESS CONTROL MEASURES

Tradit ional ly, yeast manufacturing plants have implementedincremental feed systems on the f inal fermentat ion vessels in aneffort to opt imize yeast yields and suppress ethanol formation.However, these systems were established to add a given amount ofmolasses and nutr ients over specif ied t ime intervals. Thispractice does reduce ethanol formation beyond that achieved undera total batch condit ion; however, i t does not minimize ethanolformation to the highest degree possible. A greater degree ofcontrol can be achieved by implementing a continuous monitoringsystem.

Experimental studies have shown that the ethanol productionrate is a function of the yeast growth rate, and both of theseparameters are re la ted to the res idual sugar concentrat ion.2 2 I tis therefore impor tant that the actual sugar concentrat ion in thefermentor be maintained at a low but opt imal value at al l t imes.In order to achieve th is , the fermentation process must becontinuously monitored with the aid of a computer to ant icipate

24

25

the precise demand for sugar. By continuously adding only theexact amount of molasses required by the fermentation, conditionsof excess sugar are eliminated, thus minimizing ethanolformation.

The demand for molasses depends on the cell concentration,spec i f ic growth ra te , and ce l l y ie ld . Since no sensors areavai lable that can quickly and rel iably give a direct measurementof cell mass, computer-aided material balance techniques can beused to ca lcu la te cont inuous ly the ce l l concentrat ion, spec i f icgrowth rate, sugar consumption rate, and other growth-relatedparameters.2 2 A computer can process information taken fromdirect measurement of air f low, carbon dioxide production, oxygenconsumption, and ethanol production to anticipate the demand forsugar by the system. The resul t is an ind i rect method formonitor ing yeast production that is regulated by a computer. Thecomputer continuously controls the addit ion of molasses, thereby\achieving opt imum product ivi ty with minimal ethanol product ion.However, this type of process control system is extremelyd i f f icu l t to re f ine and implement because of the t i m e delaysbetween ethanol formation in the fermentor, i ts detect ion in thestack, and the computer adjustments to the feed rates.

Another process measure that can reduce ethanol formation isin the equipment design of the aeration and mechanical agitationsystems instal led in each fermentor. The distr ibut ion of oxygenby the a i r spa rge r s y s t e m t o t he ma l t m ix tu re i s c r i t i ca l i nminimizing ethanol formation. I f oxygen is not be ing t ransferreduniformly throughout the malt, then ethanol wi l l be produced inthe oxygen-deficient areas of the fermentors. The type andpos i t ion o f baf f les and/or a h igh ly e f fec t ive mechanica l ag i ta torin the fermentors also ensures proper distr ibut ion of oxygenthroughout the malt mixture.

Fac i l i t ies that are ab le to implement feed ra te cont ro lsthrough stack gas monitoring and that have eff icient aerat ion andmixing systems can optimize yeast production and suppress ethanolformation to the highest degree. Based on available emissiontest data, i t is ant ic ipated that a reduct ion of 75 to 95 percent

26

can be achieved through the combination of feedback controls andoptimizing fermentor design.5.2 ENGINEERING CONTROLS

The most common add-on control devices for controlling VOCemissions are wet scrubbers (gas absorbers), carbon adsorbers,i nc ine ra to rs , and condensers. These types of controls are widelyused in a variety of industr ies to control VOC emissions. Thefo l lowing sect ions present a descr ip t ion of these cont ro ltechniques, operat ing parameters that affect their performance,and the feasibi l i ty of implementing these controls to reduceethanol emissions from yeast manufacturing processes. Inaddit ion to these tradit ional VOC controls, a sect ion onb io log ica l f i l t ra t ion is a lso presented be low. B io l og i ca lf i l t ra t ion is a re la t ive ly new cont ro l technique for reduc ing VOCemissions, although the engineering concept has been used forover 50 years in treating Process wastewaters.5.2.1 Wet Scrubbers (Gas Absorption)

Wet scrubbers can control ethanol emissions from yeastfermentat ion vessels. In this type of system, the contaminant isabsorbed in an absorbing liquid. Absorpt ion techniques requirelarge liquid surface areas for the incoming gas stream to makegood contact with the absorber l iquid. Good gas/ l iquid contactis increased by using hydraulic sprays, impingement trays, bubblecap trays, s ieve t rays,‘packing (modular and dump-type), grids,or a combinat ion of devices to create a high l iquid surface areawhile minimizing volume. These scrubbers can be divided into twotypes: (1) packed/tray towers or (2) spray towers.

Figure 4 presents a schematic of a packed tower. Inpacked/tray towers, the incoming gas stream enters the base ofthe scrubber and passes up through the trays or packingcountercurrent to the absorb ing l iqu id , which flows down throughthe packing or tray media. In the packing section of thescrubber, the contaminant in the gas stream is absorbed by theabsorb ing l iqu id , and the cleaned gas stream flows up and out the

27

GAS OUT

LIQUID IN

Figure 4. Schematic of packed tower absorber.

28

t op o f the un i t . The absorbing liquid flows down and out thebottom of the unit .

In spray towers, the absorbing l iquid used in the tower issprayed countercurrent to the gas flow using high-pressure spraysto create a un i form d is t r ibut ion of very smal l drop le ts o f theabsorbent within the absorber. As in packed towers, thecontaminated gas stream enters the base of the unit and flows upthe tower, where i t mixes wi th the l iqu id drop le ts in the un i t .This mixing of the gas and l iquid streams results in thecontaminant's being absorbed by the liquid. The cleaned gasstream then f lows up and out the top of the unit , and the l iquidstream f lows out the bottom of the unit .

Factors affecting the design and the performance of packedtowers include the VOC concentration and temperature of the inletgas stream, the type of absorbent used, the size and type ofpack ing mater ia l or t rays used, the l iqu id- to-gas-ra t io , thepar t icu la te load ing, and the d is t r ibut ion of the l iqu id acrossthe packing media. The first parameters given depend on theprocess being control led. The next three parameters are designparameters for the unit and depend on the type and concentrationof the contaminant to be removed. The contaminant must be highlysoluble in the absorbent selected. The part iculate loading alsoaffects the performance of the scrubber: if the gas stream has amoderate-to-high part iculate content, the packing media and spraynozzles will become clogged. Therefore, a f i l te r needs to beplaced before the absorber to-remove any particulates in the gasstream prior to entry into the absorber. The d i s t r i bu t i on o fl iquid across the packing media is cr i t ical for adequate contactbetween the contaminated gas stream and the absorbing liquid.

For spray towers, the factors affect ing performance are theconcentrat ion and temperature of the inlet gas stream, the typeof absorbent selected, the l i qu id - to -gas ra t i o , t he pa r t i cu la teloading, and the uni form d is t r ibut ion and droplet s ize of theabsorber l iqu id . The f irst four factors are the same as thosefor packed towers and affect the design of the unit in a similarmanner. The last two parameters are very important because if

29

the absorb ing l iqu id is not d is t r ibuted un i formly across theabsorber volume and the droplet sizes are too large, theperformance level of the absorber wi l l be adversely affected.

Applying wet scrubbers to control VOC emissions from yeastmanufacturing is feasible because both ethanol and acetaldehydeare extremely soluble in water. Using water as the absorbingl i q u i d , a control device eff ic iency of better than 90 percent canbe achieved. The only adverse factor associated with using ascrubber system is the amount of wastewater generated from thescrubber system and its associated treatment. The wastewatertreatment procedure used to treat scrubber eff luent should be ananaerobic treatment system that prevents or minimizes VOCemissions from the treatment process. I f t h i s t ype o f t rea tmen tsystem is not used, then the VOC emissions may not be reduced butt rans fe r red to a d i f f e ren t sou rce a t t he p lan t s i t e .5.2.2 Carbon Adsorption

Carbon adsorption has been used for the last 50 years bymany industr ies to recover a wide variety of solvents fromsolvent - laden a i r s t reams.2 3 Carbon adsorbers reduce VOCemissions by adsorbing organic compounds onto the surface ofact ivated carbon. The high surface-to-volume rat io of act ivatedcarbon and i ts preferent ia l a f f in i ty for organics make i t aneffect ive adsorbent of VOC'S. 23

The organic compounds aresubsequently desorbed from the activated carbon and recovered.

Carbon adsorbers do not apply to the low-VOC-concentrationgas streams from the final fermentation because of the lowadsorbt iv i ty o f e thanol a t leve ls less than 500 ppmv. Inadd i t i on , one carbon adsorber vendor stated that the acetaldehydepresent in the gas stream is very react ive with the carbon andwould break down the carbon, resulting in low VOC removale f f i c i e n c i e s . 2 4 Therefore, carbon adsorption was not consideredto be a v iab le cont ro l opt ion for yeast manufactur ing fac i l i t ies .5.2.3 Inc ine ra t i on

Incinerat ion is the oxidation of organic compounds byexposing the gas stream to high temperatures in the presence ofoxygen and sometimes a catalyst. Carbon dioxide and water are

30

the oxidation products. Inc inerat ion is o f ten used in indust r ieswhen solvent recovery is not economical ly feasible or pract icalsuch as at small plants or at plants using a variety of solventmixtures. 2 5 The two types of incinerators used to control VOCemission are thermal and catalyt ic. Both designs may use primaryor secondary heat recovery to reduce energy consumption.

5.2.3 .1 Thermal inc inerators . Thermal incinerators areusual ly refractory- l ined oxidat ion chambers with a burner locatedat one end. In these un i t s , pa r t o f t he so l ven t - l aden a i r i spassed through the burner along with an auxi l iary fuel. Thegases exiting the burner that are blended with the by-passedsolvent- laden air raise the temperature of the mixture to thepoint where the organics are oxidized. With most solvents,oxidizat ion occurs in less than 0.75 second at a temperature of870°C (1600°F).26

The in ter re la ted factors impor tant in inc inerator des ign andoperat ion include:

1.2.3.4.5.6.7.8.9.

10.T h e f i r s t

Type and concentration of VOC's;Solvent - laden a i r f low ra te ;Solvent - laden a i r temperature a t inc inerator in le t ;Burner type;Ef f ic iency of f lame contact (mix ing) ;Residence time;A u x i l i a r y f u e l f i r i n g r a t e ;Amount of excess air;Firebox temperature; andPreheat temperature.three parameters are characterist ics of the

fermentation process. The next three parameters arecharacter is t ics o f the des ign of the inc inerator . The aux i l ia ryfuel f i r ing rate is determined by the type and concentrat ion ofVOC's, the solvent- laden air f low rate, the f irebox temperature,and the preheat temperature. T h e a u x i l i a r y f u e l f i r i n g r a t e , t h eamount of excess air, the firebox temperature and the preheattemperature are operat ing variables that may affect theperformance of the incinerator. Well-designed and -operated

31

incinerators in industry have achieved VOC destruct ione f f i c i enc ies o f 98 pe rcen t o r be t te r . 2 5

Applying thermal incinerators to reduce VOC emissions fromthe fermentat ion process should resul t in dest ruct ionef f ic ienc ies of bet ter than 98 percent . The costs associatedwith operat ing a thermal incinerator are presented inSection 5.3.

5 . 2 . 3 . 2 C a t a l y t i c i n c i n e r a t o r s . Cata ly t ic inc inerators usea catalyst to promote the combustion of VOC's. The solvent-ladenair is preheated by a burner or heat exchanger and then broughtinto contact with the catalyst bed, where oxidation occurs.Common catalysts used are platinum or other noble metals onsupporting alumina pellets or ceramic honeycomb. Ca ta l y t i cinc inerators can ach ieve dest ruct ion e f f ic ienc ies s imi lar tothose of thermal incinerators whi le operat ing at lowertemperatures, i .e., 315° to 430°C (600° to 800°F). Thus,cata ly t ic inc inerators can operate wi th s ign i f icant ly lowerenergy costs than can thermal incinerators that do not pract ices ign i f i can t hea t recove ry . 2 6 The materials of construction mayalso be less expensive because of the lower operatingtemperatures.

Factors important in designing and operat ing catalyt icinc inerators inc lude the factors a f fect ing thermal inc ineratorsas well as the operating temperature range of the catalyst andthe presence of consti tuents in the gas stream that could foulthe ca ta l ys t . The operat ing temperature range for the catalystsets the upper VOC concentration that can be incinerated. Formost catalysts on alumina, cata lys t ac t iv i ty is severe ly reducedby exposure to temperatures greater than 700°C (1300°F). 26

Consequently, the heating value of the inlet stream must bel im i t ed . Typ i ca l l y , inlet VOC concentrations must be less than25 percent of the lower explosive l imit (LEL). The typical VOCconcentrat ions emitted from yeast manufacturing faci l i t ies areless than 25 percent of the LEL.

As with thermal incinerators, ca ta l y t i c i nc ine ra to rs a re aviable control option for reducing VOC emissions from the

32

fermentation process, hav ing t yp i ca l des t ruc t i on e f f i c i enc iesgreater than 98 percent. The costs associated with operatingcata ly t ic inc inerators are presented in Sect ion 5.3 .5.2.4 Condensation

Condensation is a process in which all or some portion ofthe volatile components in the vapor phase are transformed intothe l iquid phase. This process can be accomplished throughseveral di f ferent methods. Increasing the system pressure at aconstant temperature, reducing the temperature, or a combinationof increasing pressure and reducing temperature are possiblemethods. However, the most widely used condensation method isdecreasing the temperature at a constant pressure.

In a two-component vapor stream, where one of the componentsis noncondensable, condensation occurs when the partial pressureof the condensable component becomes equal to the component'svapor pressure. At these condit ions, the l iqu id begins to form.As the temperature of the stream is further reduced, condensationcont inues unt i l the par t ia l pressure of the vapor is equal to thevapor pressure of the liquid phase at the lower temperature. Theamount of the compound that can remain as a vapor at a giventempera tu re i s d i rec t l y re la ted to t he vo la t i l i t y o f t hecompound. The more volatile the compound, the greater the amountthat wi l l remain as a vapor. The type of coolant needed for thecondensation process depends on the temperature required forcondensation to occur.

Condensers are most effective on streams that are saturatedor nearly saturated with condensable VOC. When a gas stream isdi lute, as is the case with baker's yeast manufacturing,extens ive cool ing is requi red just to br ing the s t ream to thesa tu ra t i on po in t . Furthermore, addit ional cool ing is then neededto actually condense the VOC. Condensers are not effective ongas s t reams that conta in low-boi l ing-po in t VOC's ( i .e . , h igh lyvolatile compounds) or for gas streams that have a high flow rateof noncondensables, such as carbon dioxide, ni trogen, or air .The VOC's with low boiling points exert a high vapor pressure andhence are more dif f icult to condense total ly under normal

33

condenser operat ing condit ions. High f low ra tes o fnoncondensable gases in the stream dilute the stream and reducecondenser eff ic iency by increasing the heat load that must beremoved from the gas stream.

There are two major types of condensers: surface condensersand contact condensers. Surface condensers are usually shell andtube heat exchangers. The coolant flows through the tube, andthe vapor condenses on the outside, or shel l-side, of the tube.The condensate forms a film or masses of droplets on the tube anddra ins in to a co l lec t ion tank for s torage or d isposal . Surfacecondensers usually require more auxiliary equipment than contactcondensers, but solvent recovery is possible since the coolantand the condensate are kept separate. Also, the coolant cannotbecome contaminated in a surface condenser. Equipment typical lyneeded for surface condensers includes dehumidificationequipment, a shel l -and- tube heat exchanger , a re f r igerat ion uni t ,a recovery tank for the condensate, and a pump to dischargerecovered VOC to storage or disposal.

Contact condensers cool the vapor by spraying a liquid,usual ly a t ambient temperature or s l ight ly ch i l led, in to the gasstream. The result is int imate mixing of the gas stream and thecooling medium. In some instances, contact condensers act asscrubbers in that they col lect noncondensables that are misciblewith the cooling medium. Contact condensers are simple in designand re la t i ve l y i nexpens i ve to i ns ta l l . Although contactcondensers have advantages, the i r app l ica t ion is l im i ted because,l ike wet scrubbers, the VOC-contaminated coolant cannot normallybe reused direct ly.

The most obvious area to use condensers in the baker's yeastmanufacturing process is during the early stages of yeast growth(pure cu l ture s tages) . In these ear ly s tages, the concentrat ionof ethanol is expected to be at i ts highest and the airf low ratesare lowest. Although no emissions test data were available,in format ion suppl ied f rom yeast manufactur ing fac i l i t iesind icates that the a i r f low rate f rom pure cu l ture fermentat ion istypically 10 m3/min (400 ft3/min) with an average VOC

34

concentration of 5,000 to 10,000 ppmv. At these leve ls ,condensers should achieve better than 90 percent emissionreduction at condensation temperatures below -40°C (-40°F).5 . 2 . 5 B i o l o g i c a l F i l t r a t i o n

Figure 5 presents a s impl i f ied schemat ic o f a b io f i l t ra t ionsystem. A b io log ica l f i l te r is bas ica l ly a compost bed that hasbeen inoculated with aerobic microorganisms. T h e f i l t e reliminates VOC emissions by passing the VOC-laden gas streamthrough the compost bed. As the gas stream passes through thebed, the contaminants are removed through adsorption, absorption,and chemical degradation. Port ions of the contaminants areadsorbed by the compost material while others are absorbed by thewater in the bed. These contaminants are then metabolized by themicroorganisms and are converted to carbon dioxide and water.Ethanol conversion, in the case of yeast fermentat ion, isaccomplished in two stages. F i r s t , one type of microorganismconsumes the alcohol and converts it to organic acids. In thesecond stage, another microorganism converts the organic acidsinto carbon dioxide and water. A delicate balance between thetwo microorganisms must be maintained to ensure proper operationo f t h e b i o l o g i c a l f i l t e r s . After the gas stream passes throughthe bed, the cleaned gas is vented out stacks located at the topo f t h e f i l t r a t i o n u n i t . Volat i le organic compound eff ic ienciesof better than 90 percent have been obtained when biof i l t rat ionunits have been instal led to control emissions from yeastfermentat ion vessels.2 7

The cr i t ica l parameters for a b io f i l t ra t ion system are theVOC concentration in the inlet gas stream, the pH and moisturecontent of the bed, and the bed temperature. The bed temperaturemust be maintained above 10°C (50°F) or the microorganisms in thecompost will become dormant. The pH and moisture content of thebed should be in the range of 6.5 to 7.0 and 65 to 70 percent,respect ive ly . Water spray l ines are located at the top of theb io f i l t r a t i on un i t t o he lp ma in ta in the mo is tu re con ten t a t t heappropr ia te leve l . The in le t VOC concent ra t ion is a lso cr i t ica lto the per formance of the b io f i l t ra t ion system. The

35

Figure 5. Schematic of a biofi l tration system.

biof i l t rat ion system is designed to handle di lute VOC streamswi th f ixed concentrat ions. F luctuat ions in the concentrat ion orhigh VOC concentrations result in an imbalance between the twocontrol l ing microorganisms. The microorganism that convertsalcohols to organic acids assimi lates the alcohols at a rate thatdecreases the pH of the bed, making the bed acidic. A low pHwi l l k i l l the o ther type o f microorganism, the bed wi l l no longerope ra te e f f i c i en t l y , and incomplete conversion of the alcoholsw i l l occu r . For th is reason, b io f i l t ra t ion systems were notdesigned for batch processes but for continuous, dilute VOCstreams. However, a cont ro l system ( i .e . , wet scrubber) locatedupstream of the biof i l t rat ion system that could control the peakconcentrat ion levels would result in a fair ly constant di lute VOCstream to the b io f i l t ra t ion system. This combination of ascrubber plus a biof i l t rat ion system would be a pract icalapproach to controlling emissions from a batch process.5.3 IMPACT ASSESSMENT OF CONTROL OPTIONS

Based on the technical evaluation of traditional add-on VOCcontro l techniques; the most promising opt ions for control l ingVOC emissions from the final fermentations appear to be wetscrubbers, thermal inc inerators , and ca ta l y t i c i nc ine ra to rs .Therefore, these control systems were evaluated further todetermine their associated cost and environmental impacts. Theimpacts of the control options were determined based on theeffectiveness of the control systems to handle the model emissionstream presented in Section 4.4.

Table 5 summarizes the cost and environmental impacts ofeach control system. To determine the cost impacts of thecont ro l opt ions, cost algorithms were developed based on standardE P A m e t h o d s . 28,29

Tables 6 and 7 give details of the annual andcapital costs associated with each control device. Theenvironmental impacts were derived as a function of the designand size of the control system. As shown in Table 5, catalyt icincinerators appear to be the most cost-effect ive control systemfo r t h i s app l i ca t i on .

37

*Source: OAQPS Control Cost Manual. Fourth Edition. EPA 450/3-90-006. January 1990. Chapter 3.**Source: Organic Chemical Manufacturing Volume 5: Adsorption, Condensation, and Absorption Devices. EPA Report

No. 450/3-80-027. December 1980. Appendix B.

40

The cost of control could be reduced further, however, bythe use of a combination of control systems. A process using awet scrubber fo l lowed by an inc inerator or a b io log ica l f i l te rcould be conceptual ized with the expectat ion that the annual costwould be reduced. The water from the scrubber would not beconsidered "wastewater" but would be recycled within the processand used to continually generate an emission stream with aconstant concentrat ion of ethanol at a reduced air volume. Thegeneration of a low-volume constant ethanol concentration streamwould allow the use of a smaller incinerator system or the use ofa b io l og i ca l f i l t r a t i on sys tem, which could conceivably result inlower control costs. Because of the complexity of such a systemand the need to consider plant specif ic condit ions, a costanalys is is not inc luded in th is repor t .

I n add i t i on , impacts for condensers were also evaluatedbased on the control of emissions from the pure culturefermentat ions. Table 8 gives detai ls on the cost ing associatedwith using condensers for control l ing emissions from the purecul ture fermentors. These costs were developed based on atypical air f low rate from pure culture fermentat ion of 10 m3 /min(400 ft3/min) at a VOC concentration of 7,500 ppmv. Based on a95 percent control requirement, the condensation temperature ofthe condenser would be -47°C (-53ºF). The capital cost perfermentor is estimated at $250,000. The annual operating cost isestimated at $111,000/yr. Based on an emission reduction of27 Mg/yr (29 tons/yr), the average cost -e f fect iveness forcondensers is $4,200/Mg ($3,800/ton).6.0 CONCLUSIONS

The typical yeast manufactur ing faci l i ty emits approximately82 Mg/yr (90 tons/yr) of VOC emissions. The primary consti tuentsin the emission stream are ethanol and acetaldehyde, with ethanolcomprising approximately 80 to 90 percent of the emissions andacetaldehyde comprising the remaining 10 to 20 percent ofemissions. The primary emission sources are the final tradefermentat ions, which account for 80 to 90 percent of the totalf ac i l i t y em iss ions .

41

*Source: OAQPS Control Cost Manual.. Fourth Edition. EPA 450/3-90-006. January 1990. Draft Chapter.

42

The two types of control measures that are currentlyemployed at yeast manufactur ing faci l i t ies are (1) processcontrol and (2) add-on controls. The majori ty of yeastmanufacturers use a moderate degree of process control in thef inal fermentat ion stages to reduce ethanol formation. However,these process control measures can be enhanced by implementingcomputer-based feed rate controls and improving fermentordesigns. Implementing a computer-based feed rate control systemand improved fermentor design can potentially suppress ethanolformation by 75 to 95 percent.

One yeast manufacturer has applied a combination wetscrubber and biof i l t rat ion system for control l ing VOC emissions.Performance data from this unit suggests an emission controle f f ic iency of bet ter than 90 percent . 2 7 Other add-on controltechniques that could potent ial ly be appl ied to the yeastfermentat ion process are incinerators. The control technologyevaluat ion suggests that a cata ly t ic inc inerator is the mostcost-effect ive approach for reducing VOC emissions, i f the use ofa single control device is appl ied to the emission stream.However, a combination system such as that described above or acombination of a wet scrubber and incinerator could result inlower control costs and relat ively equivalent emissionreductions.

At present, no process control measures or add-on pollutioncontrol systems are currently being used to reduce VOC emissionsfrom the pure culture fermentations. However, based on thein format ion suppl ied by yeast manufactur ing fac i l i t ies , i tappears that adding process control measures to the pure culturefermentations or applying condensers could potentially reduce VOCemissions from this stage of the process by better than90 percent.

43

7.0 REFERENCES

1. Clean Air-Ozone. Press release, Associated Press. May 3,1988.

2. Chen, S. L., and M. Chiger. Production of Baker's Yeast.Comprehensive Biotechnology. Vol. 20. NY, Pergamon Press. p.430.

3. Reed, G., and H. Peppler. Yeast Technology. Wesport, CT, AviPublishing Company. 1973. pp. 54-55.

4. Reference 2, p. 433.

5. Reference 3, p. 58.

6. Reference 3, pp. 57-58.

7. Reference 3, p. 60.

8. Reference 2, pp. 441-442.

9. Reference 2, p. 442.

10. Reference 2, pp. 74-75.

11. Reference 2, p. 80.

12. Reference 2, p. 84-85.

13. Reference 2, p. 90.

14. Reference 2, p. 91-92.

15. Reference 2, p. 96.

16. Industrial Wastewater Volatile Organic Emissions--BackgroundInformation for BACT/LAER Determinations. Contro lTechnology Center. U. S. Environmental Protection Agency,Research Triangle Park, NC. EPA No. 450/3-90-004. January1990.

17. Fermentor Emissions Test Report. Universal FoodsCorporation, Baltimore, MD. Prepared by Gannett Fleming,I n c . , Baltimore, MD. October 1990.

18. Fermentor Emissions Test Report. Gist-brocades *FoodIngredients, Inc., East Brunswick, NJ. Prepared by TraceTechnologies, Inc., Bridgewater, NJ. Apri l 1989.

19. Fermentor Emissions Test Report. American Yeast, Baltimore,MD. Prepared by Gannett Fleming, Inc., Baltimore, MD.August 1990.

44

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

Fermentor Emissions Test Report. Universal Foods, Inc.,Baltimore, MD. Prepared by Universal Foods, Inc.,Milwaukee, WI. 1990.

Fermentor Emissions Test Report. Universal Foods, Inc.,Baltimore, MD. Prepared by Environmental Technology andEngineering Corporation, Elm Grove, WI. May 1989.

Wang, H. Y., .et al. Computer Control of Baker's YeastProduction. Biotechnology and Bioengineering, Volume 21.1979. pp. 975-995.

U. S. Environmental Protection Agency. Polymeric Coating ofSupporting Substrates - Background Information for ProposedStandards. EPA-450/3-85-022a. Research, Triangle Park, NC.Apri l 1987. p. 4-10.

Telecon. Hale, C., MRI, to Wersal, D., VIC Manufacturing,Minneapolis, MN. September 3, 1990. Information regardingthe use of carbon adsorbers for controlling VOC emissionsfrom yeast fermentation processes.

Reference 23, p. 4-30.

Reference 23, p. 4-31.

Memo from Barker, R., and M. Williamson, MRI, to Smith, M.,ESD/CPB. Trip report for Gist-brocades, East Brunswick, NJ.August 1991.

Control Technologies for Hazardous Air Pol lutants. U. S.Environmental Protection Agency, Research Triangle Park, NC.EPA/625/6-86/014. September 1986.

OAQPS Control Cost-Manual. U. S. Environmental ProtectionAgency , Research Triangle Park, NC. EPA 450/3-90-006.January 1990.

45