Документ взят из кэша поисковой машины. Адрес оригинального документа : http://www.enzyme.chem.msu.ru/ekbio/article/comb.pdf Дата изменения: Wed Mar 30 13:55:28 2005 Дата индексирования: Mon Oct 1 21:47:03 2012 Кодировка:

~

.
Combinedbiological and physico-chemicaltreatment of
baker's yeast wastewater including removal of coloured and recalcitrant to biodegradation pollutants
M. Gladchenko., E. Starostina.., s. Shcherbakov.., B. Versprille... and S. Kalyuzhnyi.

~

~
(J)
OJ 0 (!)

~.
~ c-

. DepartmentChemical of Enzymology, Chemistry Faculty, Moscow State University, 119992 Moscow,
Russia(E-mail: mag@enzyme.chem.msu.ru) .. Department Grape ProcessingTechnology,MoscowStateUniversity FoodIndustry, of of
Volokolamskoye shosse 11, 125080 Moscow, Russia

~
0

~
0co
'< <

...

Biothane

Systems

International,

Tanthofdreef

21,2623

EW

Delft,

The

Netherlands

(E-mail: bram.versprille@biothane.n/) Abstract The UASBreactor(35°C)was quiteefficientfor removal bulkCOD (62-67%) evenfor such of
high strength and recalcitrant wastewater as the cultivation medium from the first separation process of

~ ~
0 ~ ... .!..

2U1

baker'syeasts(theaverage organicloadingratesvariedfrom 3.7 to 10.3 g COD/lId). Theaerobic-anoxic biofilter(20°C)can be usedfor removal remaining of BOD andammonia from strongnitrogenous anaerobic effluents;however,it sufferedfrom COD-deficiency fulfil denitrification to requirements. balancethe To CODIN ratio,somebypassof raw wastewatershouldbe addedto the biofilterfeed.Theapplication iron of chloridecoagulation post-treatment aerobiceffluentsmayfulfilthe dischargelimits(evenfor colour for of mainly exertedby hardlybiodegradable melanoidins) underironconcentrations around200 mg/l. Keywords Aerobic-anoxic biofilter;b~ker'syeastwastewater;ironcoagulation; melanoid UASBreactor ins;

@

~
~ "1J §-

~
~.
'" 0

Introduction
The baker's yeast industry is very popular in Russia. Such factories exist in almost all Russian provinces and altogether they produced 56 mIn. m3 wastewater per year (Kalyuzhnyi et al., 2003). Since the main substratefor baker's yeastproduction in Russiais sugar beet molasses, these wastewaters are high strength (10-80 g COD/l) , strong nitrogenous(0.5-1.5 g/l total N), sulphate-rich (2-10 g/l), phosphorusvariable (sometimes P-deficient), recalcitrant for biodegradation and highly coloured (melanoidins etc.) ones. Currently many yeast factories are faced with heavy trade-effluent charges.Land disposal options generateproblems with ground water pollution and areprohibited in majority of the Russian regions. Many local municipal sewage treatment plants are now insisting on pre-treatmentof such effluents before discharge into their sewerage.The objective of this paper was to develop a lab scale technology for treatment of baker's yeast wastewater to meetthe limits for dischargeof treatedwastewaterinto municipal sewerage. The most troublesome limits in this case are the following (mg/l, except colour): COD - 800; SO42-500; total N -100; N-NH3~50; P-PO43--3.5; colour threshold 1:128 (optical density < 0.2 at dominant wavelength). As a first treatment step,the UASB reactor operating at 35°C was applied for the elimination of the major part of COD andconcomitant sulphatereduction. In a subsequent step,the biofilter operating in alternative aerobic-anoxic regime at -20°C was used for the removal of the remaining BOD and nitrogen. Finally, iron coagulation was applied to fulfil the limits on COD, PO43-and colour. Materials and methods Wastewater Sincethe major (-30% of total volume wastewaterproduced) and the strongeststreamfrom

~

67


~ Q
po

g. ~
~

~

~

.

yeast factories (Kalyuzhnyi et al., 2003) is the cultivation medium obtained after the first separation of yeasts (CM-IS), this stream taken from a Moscow baker's yeast factory is used in this study. The other reasonis that this streamusually shows less variation in composition than the generaleffluent of yeastfactories. Somecharacteristicsof this wastewater during the period of sampling from November, 2002 to March, 2003 are given in Table 1. Gel-filtration of untreatedWW (Figure la) revealedthat the substances responsiblefor visible colour (ODS80) have symmetrical Gauss-typedistribution of their molecular weights (MW) in the rangeof 0-30 kDa with a maximum at 6-7.8 ilia (>60% of colour is associated with the substanceshaving MWs in the range of 2.6-14 ilia). On the contrary, substanceswith aromatic structures (OD280)have non-symmetrical distribution of their MW s

(Figure la) with a clearprevalence low MW molecules of (maximumat 2.6-4.5ilia; >
82% ofUV-absorption is associatedwith MWs < 8 ilia). Thesedata show that the visible colour yeast wastewateris not closely associatedwith aromatics like phenolic compounds.
Gel-filtration

The gel SephadexG-50 (Pharmacia, Sweden) was equilibrated with 0;1 N NaCl (pH 7.6) that was further applied as a mobile phasein a gel-filtration column. Potassiumbichromate (fully permeating into gel) and blue dextran (fully non-permeating) were used for the calibration of the column. Basedon elution volume (V e) of calibration substances, folthe lowing equation for determination of MW (Ve of bichromate and blue dextrane were assignedto column applicability limits - 1,500 and 30,000 Da, respectively) was obtained:
Table 1 Range of variation of some characteristics of the CM-1 S (average values from 5 samplings are given in brackets)
CODtot' g/l

.

17.9-31.1 (22.5)
Total N, mg/l 993-1,651 (1,179) Phenolic compounds

713-1,167(875)

CODss' g/l 0.95-2.93 (1.07) N-NH3, mg/l 186-450(278) Dominant wavelength,nm 580

CODcol' g/l 0.82-1.91 (1.46) Total P, mg/l 12-78(32) Colour purity, % 45.7-52.8(48.6)

CODsol' g/l 15.0-26.6 (19.0) P-P04, mg/l 2-32(9) Colour luminance, % 40.6-50.2(46.6)

pH 4.01-5.68 (5.14) 504, mg/l 682-3,028(1,828) ODS8o 0.76-1.17(0.95)

.

80D580

15 12
9

80D580

15 12
9 6

0 OD280

i
'is
~

0 OD280

i
.~

!~

~ ~
~

3 0 I>~ 'I/"'~ ~~ ~. .",~ ~ "I)'" '+I'""' "1~1)

I>~ 'I/",b ~~ ~.

.",~

':-'~ "I)"

'+I'""'"1~

MWrange. kD. 15 8 OD580 0 0D280 12
9

MWrange, kD. 20

i
.~

16
12

oS
.~

~ 6!3
~ I> b '1/"" ~ 'b .",. ~ ~ ':-' ~ ~ -" .:,'-

~ 8!4
~ I> b '1/"" ~ 'b .",. ~ ~ ,:-," -'" ~v -" .:,""

~

~

0

II

~.

"~,,

~.

"I)"

MWrange. kD. 68

MWrange, kD.

Figure 1 MW distribution of substances exerting light absorption at 580 and 280 nm in raw CM-1 S (a), anaerobic (b), aerobic (c) and final (d) effluents


MW = exp(12.521-O.1303Ve).During gel-filtration of untreated and treated wastewater,
, the eluting fractions (2.1 ml) were collected and their optical densities at 580 (dominant wavelength) and 280 (characteristic wavelength for aromatics) nm were determined.
UASB reactor

Laboratory UASBreactor (rectangular cross-section cm2, 33.64 height 82cm,totalworking volume 2.731)made fromtransparent plasticandequipped 6 sampling with portsalongthe reactor height used. reactor keptin thethermostat was The was (under :t 1°C)andwasseed35 edwith granular sludge (66.5g VSS,specific aceticlastic activity- 0.3g COD/gVSS/day at
30°C) from the full-scale EGSB reactor treating brewery wastewater (Efes-Moscow). Assessment sludgeaceticlasticactivities anddeterminationof anaerobic of biodegradabilityof the CM-lS wereperformedasdescribedpreviously (GladchenkoandKalyuzhnyi, 2003).
Anaerobic-anoxic bifllter

~ Q '" g~

~ 7'"

~

~

The tubular biofilter (diameter - 5 cm,height- 55 cm,packed 0.5-2 cmfractionof road by metal) had the working volume of 0.71 and functioned in alternating aerobic/anoxic regime for treatment of the anaerobiceffluents under ambient temperaturein the laboratory (20 :t 3°C). The operation schemeincluded a sequencingprocesswith a one-hour cycle consisting of 4 phases.During the first unfed phase, air at a flow rate of 0.8 l/min was pumped through an external loop of the biofilter. Aeration was switched off throughout the second unfed phasewhile the high recycle rate of effluent (0.1251/min) was applied to ensure an adequatemixing and a complete consumption of the remaining soluble oxygen in the biofilter. During these 2 phases,nitrification and oxidation of remaining BOD proceeded. Then the feeding was performed during the 3rd phaseunder the samerecycle rate of effluent. The last phaseincluded only'mixing (by effluent recycle) and was variable to close the 1 h working cycle of the programmable multi-channel timer controlling all 3 (air, recycle, feeding) pumps used. During the last 2 phases,denitrification proceeded.In the middle of the external loop of biofilter, an electronic sensor was inserted for on-line monitoring of soluble oxygen. The attached nitrifying-denitrifying biomass formed in the biofilter during the previous research (Gladchenko and Kalyuzhnyi, 2003) was directly used for treatment of anaerobiceffluents in this study. The excessof sludge was periodically withdrawn by intensive backwashofbiofilter. Coagulation assays They were performed with 200 ml of biofilter effluent in a laboratory glass under continuous stirring and pH control. Addition of coagulant (FeC13'6H20)was carried out under 200 rpm, then intensity of stirring was reduced to 40 rpm to complete a flocculation processduring which pH was maintained at 7.2-7.5 by addition of sodium hydroxide. Analyses All analyses were performed by Standard Methods (1995) or as described previously (Gladchenko and Kalyuzhnyi, 2003). All gas measurements were recalculated to standard conditions (1 atm, O°C).Statistical analysis of data was performed using Microsoft ;Excel. Results and discussion
UASB reactor performance

In the preliminary experiments, it wasfound that the raw CM-lS's were quite biodegradable in anaerobic conditions (>90% on COD basis). Some results of the mesophilic UASB treatment (35°C) of the raw CM-lS's under quasi-steadystateoperation are shown in Table 2 (UASB runs). It can be seenthat a stepwisedecreaseof hydraulic retention time

69


~

(HRT) from 4.9 to 2 days (organic loading rate (OLR) finally exceeded 10 g COD/l/d) almost did not influence total COD removal, which was in the range of 62-67%. Only tracesofVFA were detectedin the effluents (data not shown). However, suchexhaustionof easily biodegradable COD in the anaerobic effluents might create COD deficiency problems for subsequentbiological nitrogen removal. In spite of acidic influent pH feed, the effluent pH was close to 8 as a result of VFA consumption and mineralisation of nitroge-

nousspecies ammonia to (Table2, UASBruns).Theconcentrations phosphate of increased
(except run UASB-3) in the effluents (Table 2) due to mineralisation of phosphoric species. On the contrary, a gradual development of biological sulphate reduction led to an almost complete disappearanceof sulphate in the UASB reactor (Table 2). The latter is almost quantitatively recoveredas soluble sulphide in the sludge blanket zone of the reactor (data not shown). However, sulphide concentrationsin the collected effluents were significantly lower due to its oxidation occurring in the settler zone of the UASB reactor and effluent collection vessel- the whitish precipitates (presumably, elemental sulphur) were clearly seen on the reactor and settler walls. Such lossesoccuued due to the small size of the laboratory reactor are hardly possible in full-scale industrial reactors. Colour removal was generally insignificant during this stage(Table 2). Gel-filtration of anaerobiceffluent (Figure 1b) also did not reveal significant changesin MW distribution of coloured and aromatic substancescompared to influent (Figure 1a) indicating that these substancesare persistent in anaerobic conditions. Throughout runs 1-3 (70 days), the quantity of sludge inside the UASB reactor almost doubled, its specific aceticlastic activity slightly increasedand the VSS/TSS ratio slightly decreased (Table 3).
Performance of alternative aerobic-anoxic biofllter

Q
~

~
~

~ ~

~

In order to simulate a possible scenario for a yeast factory, when only the strongest stream (CM-1S) is treated anaerobically, then the effluent is mixed with the other less
Table 2 Operational parameters and efficiency of the biological reactors (mean :t standard deviation) Parameter/run HR7; days OLR,gCODto/l/d Influent CODtot' g/l Effluent CODtot' g/l Total COD removal, % Influent pH Effluent pH Influent Ntot' mg/l Effluent Ntot' mg/l Total N removal, % InfluentN-NH3,mg/l Effluent N-NH3' mg/l N-NH3 removal, % Effluent N-N03, mg/l Effluent N-N02, mg/l UASB-1 4.86:tO.08 4.67:tO.03 22.33 7.57:t1.07 66.1:t4.8 5.68 8.16:tO.11 993 ND UASB-2 3. 28:tO.03 5.50:t0.01 17.93 6.63:tO.62 62.2:t2.9 5.61 8.17:t0.09 998 ND UASB.3 1.98:tO.11 10.26:tO.61 20.25 6. 72:tO. 12 66.8:tO.6 4.99 7.95:tO.10 1075 ND AeAnB-1 3.11 :to.07 1.16:tO.16 3.59:tO.48 1. 15:tO.17 68.1:t1.7 8.14:tO.18 7.83:tO.1 474:t10 290:t7 AeAnB-2' 4.35:tO.44 1.73:tO.03 7.9:tO.04 2.30:t0.02 71.0:t0.2 7.39:tO.07 7.63:tO.12 636:t11 14O:t5

202 702:t25 0 0

186 783:t12 0 0

235 729:t30 0 0

38.8:t3.5
416:t21 25:t5 94.0:t0.9 207:t31 traces

78.O:t1.0
558:t3 64:t5 88.5:t1.0 12:t2 traces

Denitrificationefficien.,%
Influent P-PO 4' mg/l Effluent P-P04' mg/l Influent504,mg/l Effluent 5°4, mg/l Influent OD580 EffluentODseo ODseo removal, %

2.3 11.1:t1.6 3028 448:t76 1.11 :to.09 0.86:tO.14 22.4:t6.4

1.8 4.8:t1.5 1917 41 :t25 0.76 0.735 3.3

5.6 4.2:tO.1 1245 traces 0.755 0.73 3.3

47.1:t6.5*
3.2:tO.8 3.1:t1.2 17:t15 246:t35 0.38:tO.01 0.33:tO.01 11.2:tO.9

86.4:t1.1*
7. 7:tO.1 4.1:tO.8 164:t3 453:t41 0.53:tO.01 0.41:tO.01 22.5:t1.2

*Calculated as: [1 - {[N-NO~ef + [N-N02IeJ/([N-NH3Iin - [N-NH3]eJI*100 #11 % of raw CM-1 5 were added to balance COD/N ratio for denitrification
70 ND

-

not determined


.
Table 3 Some sludge characteristics of a mesophilic UASB reactor treating the raw CM-1 S. Parameter VSS in the reactor, 9 TSSinthereactor,g VSS/TSS, % Aceticlastic activity, 9 COD/g VSS/day Start 66.5 101.6 65.5 0.30 End 115.6 181.2 63.8 0.35

~ G) 5"

concentratedfactory wastewatersand the UASB effluents were diluted (in -2 times by tap water) before being fed to biofilter. After tuning of durations of the aerobic and anoxic phases,the following results were obtained (Table 2, run AeAnB-l). It is seenthat the averagetotal COD and ammoniaremovals accountedfor 68% and 94%, respectively. However, the effluent nitrate concentrations were relatively high (207 mg N/l, on the average) that was related with COD deficiency to have a stable denitrification - somepart of the incoming COD (-1.15 g/l, Table 2, run AeAnB-1) was non-biodegradable. To balance the COD/N ratio, somebypassof raw CM-lS (11%) was addedto the biofilterfeed during run AeAnB-2. This led to a substantialdecrease nitrate but someincreaseof ammonia in aerof obic effluents (Table 2, run AeAnB-2). It seemsthat it is hardly possible to reach a lower level of ammonia in the effluent due to an immanent drawback of this relatively simple biofilter construction where wastewater filling and effluent withdrawal were performed simultaneously in a CSTR regime. Though during run AeAnB-2 the total inorganic nitrogen concentrationswere around 76 mg N/l (Table 2), the aerobic effluents contained also significant concentrations (65 mg N/l) of organic nitrogen (seemsto be hardly biodegradable) resulting in total nitrogen concentrationsaround 140 mg N/l, i.e., higher than the discharge limit to sewer (100 mg' N/l). The total COD concentrations during run AeAnB-2 were close to the biodegradability limit of yeast WW but higher (Table 2) than discharge limit to sewer (800 mg/l). Due to oxidation of sulphide presentedin anaerobic effluents, sulphateconcentrations were below the discharge limit (500 mg/l); however, this was due to elementary sulphur pre-settling before biofilter treatment. Since such pre-settling may be difficult in implementation in full-scale conditions, sulphatecan be a concernfor biologically treated baker's yeast wastewater. The phosphate concentrations (Table 2, run AeAnB-2) in the aerobic-anoxic effluents were close to the dischargelimit (3.5 mg P/l). In spite of 63% removal of phenolic compounds (data not shown) during the aerobic-anoxic stage,colour removal accountedfor only 23% (Table 2, run AeAnB-2). This is in accordancewith gel-filtration data for biofilter effluents (Figure lc) when a shift of maximum of MW distribution of coloured substances the range of 4.5-6 kDa and a decrease relative to of content of low MW substances 2 kDa) was observedcompared to untreated(Figure la) « or anaerobically treated (Figure lb) yeast wastewater. Thus, the visible colour is mainly associated with the other substances than phenolic compounds(e.g., persistentto biodegradation melanoids according to literature data (FranciscaKalavathi et al., 2001».
Performance of Iron coagulation step

g:

~ ~
~

Some results on efficiency of the iron coagulation step for treatment of mixed aerobicanoxic effluent are presentedin Table 4. It is seenthat all targeted parameters(total COD and nitrogen, phosphate,ammonia and colour) decreased with increasing acting Fe concentrations and the discharge limits are already achievable under iron concentrations around 200 mg/l. The colour of wastewaterurtderwentdramatic changesfrom deepbrown to pastel yellow after coagulation with 196mg Fe/l. The gel filtration data for final effluent obtained (Figure 1d) also showed a substantialrelative decreaseof all fractions with MW < 18 ilia (colour-bearing fractions). A relative increaseof fractions with MW > 18 ilia was due to

71


Table 4 Performance ironcoagulation of step ActingFeconcentration, mg/l 196 476 800 107 690 96

0 COOtol,mg/l Phenols, mg/l 1,310 234

tg 1,110 208

90t 550 50

1304 400 44

~ (;, [00580 §-

P-P04,mg/l
TotalN N-NH3' mg/l Sludge percentage, % vol'.

4.7
120 34 0.325 -

3.1
101 25 0.298 NO

1.4
82 16 0.170 13.7

0.1
63 10 0.071 59.1

traces
34 5 0.021 89.1

traces
25 3 0.014 90.4

~ ~
~

SVI, ml/g TSS' SludgeVSS/TSS,%'

-

NO NO

259 89.6

440 33.5

397 31.2

295 25.0

.

' after minof settling NO notdetermined - 30 formation of iron-induced aggregateswith higher MWs having poor settling properties. These results are superior (with regard to coagulant added) to those reported in the literature for anaerobically treated baker's yeast wastewater(Kalyuzhnyi et al., 2003). It is likely that the additional removal of COD (and partly colour) occurring on the aerobic-anoxic step led to a significant economy in coagulant addition. The sludge formed under an acting Fe concentration of 196 mg/l was relatively large (SVI = 259 ml/g TSS) and had high (-90%) VSS content (Table 2) showing the significant removal of organic COD and nitrogen during the iron coagulation step. Conclusions 1. The VASB reactor is quite efficient for removal of bulk COD (62-67% efficiency) even for suchhigh strength wastewateras cultivation medium obtained after the first separation of yeasts. 2. The aerobic-anoxic biofiltercan be usedfor removal of the remaining BOD and ammonia from anaerobiceffluents; however, it suffered from COD-deficiency to fulfill denitrification requirements. To balance COD/N ratio, some bypass of raw wastewater should be addedto the biofilter feed. 3. The application of iron chloride coagulation for post-treatment of aerobic-anoxic effluents may fulfill the dischargelimits to the sewer under iron concentrationsaround 200 mg/!. Acknowledgements The financial support of Biothane SystemsInternational is gratefully acknowledged. We thank the Moscow baker's yeast factory for delivering the CM -1S. References
Francisca Kalavathi, D., Uma, L. and Subramanian, G. (2001). Degradation and metabolization of the pigment-melanoidin in distillery effluent by the marine cyanobacterium Oscillatoria boryana bdu 92. Enzyme Microb. Technol., 29, 246-251. Gladchenko, M. and Kalyuzhnyi, S. (2003). Development of the energy efficient technology for treatment of the high strength and strong nitrogenous landfilileachates. Engineering & Environment Protection, 6(1), 107-119. Kalyuzhnyi, S.Y., Gladchenko, M.A., Starostina, Ye.A., Shcherbakov, S.S. and Korthout, D. (2003). Highrate anaerobic treatment as a key step of purification of baker's yeast wastewater: a review. Manufacture of alcohol and liqueur & vodka products, N3, 37-44. Standard Methods for the Examination of Water and Wastewater (1995). 19th edn, American Public Health Association! American Water Works Association/W ater Environment Federation, Washington DC,

72

USA.