hal,1-9(sumihar hds)

JURNAL ITENAS KINETIC STUDY ON THE TREATMENT OF SYNTHETIC PALM OIL MILL EFFLUENT BY ANAEROBIC MEMBRANE BIOREACTOR (AMB) Sumihar H.D.S 1) T. Setiadi 2) , I.G. Wenten 2) 1)

Department of Chemical Engineering, Institute of Technology Nasional Bandung 2) Department of Chemical Engineering, Institute of Technology Bandung ABSTRACT

In this study, a membrane coupled with an anaerobic bioreactor was used to treat synthetic palm oil mill effluent (POME). A hollow fiber microfilration (MF) membrane module was used to filter final effluent and simultaneously retain biomass in bioreactor. The porous diameter of membrane was 0.2mm and surface area of 0.04 m2. The applied Hydraulic Retention Time (HRT) was 1 and 2.5 days and Solid Retention Time (SRT) was 25, 35, 45, and 55 days. From this study was found that flux were in the range of 12 to 16 L/m2.hr, reactor mixed liquor suspended solids (MLSS) between 2000 and 25,000 mg/L and COD removal percentage was between 89 and 91 %. Methane composition, produced during this process, ranged from 5-14 %. Kinetic parameters depended on HRT. At HRT of 1 day, the maximum specific growth (mm), yield (Y), specific decay rate (kd) and constant half saturation (Ks) were 2.25 .10-2 day-1, 1.22 .10-3 mg COD/mg MLSS, 2.03.10-2 day-1 -3 -2 -1 -3 and 2.52 .10 mg/L, respectively. At HRT of 2.5 days, mm, Y, kd, and Ks were 1.3.10 day , 3.92 .10 mg -2 -1 3 COD/mg MLSS, 5.83 .10 day and 2.05 .10 mg/L, respectively. Key words: anaerobic, pome, membrane bioreactor, kinetic parameters.

ABSTRAK Pada penelitian ini, membran dikombinasikan dengan bioreaktor anaerob untuk mengolah limbah cair minyak sawit sintetik. Membran mikrofiltrasi dengan modul buluh berongga digunakan untuk menyaring luaran dan sekaligus menahan biomassa tidak terbawa luaran. Diameter pori membran yang digunakan adalah 0.2m, luas permukaan 0.04 m2. Waktu tinggal cairan (WTC) yang digunakan adalah 1 dan 2.5 hari dengan Waktu Tinggal Padatan (WTP) 25, 35, 45 dan 55 hari. Dari hasil penelitian yang dilakukan, diperoleh fluks pada rentang 12-16 L/m2.jam, konsentrasi biomassa (MLSS) antara 2000 - 25.000 mg/L dan persentase penyisihan COD antara 89 91 %. Komposisi metana yang diperoleh yaitu 5-14 %. Parameter kinetik yang diperoleh menunjukkan ketergantungan pada waktu tinggal cairan (WTC). Pada WTC 1 hari, laju pertumbuhan maksimum spesifik (mm), perolehan (Y), laju kematian spesifik (kd) and konstanta setengah jenuh (Ks) adalah 2.25 .10-2 hari-1, 1.22 .10-3 mg COD/mg MLSS, 2.03 .10-2 hari-1 dan 2.52 .10-3 mg/L. Pada WTC 2.5 hari mm, Y, kd, dan Ks adalah 1.3 .10-2 hari-1, 3.92 .10-3 mg COD/mg MLSS, 5.83 .10-2 hari-1 dan 2.05 .103 mg/L. Kata kunci: anaerobik, luaran pabrik minyak sawit, bioreaktor membran, parameter kinetik.

No. 1, Vol. 6, Maret - Mei 2002

1

JURNAL ITENAS INTRODUCTION Recently, the world demand of crude palm oil is continuously increasing. In 2005 Indonesia is predicted to produce 10 million tons of crude palm oil (CPO). Every ton of CPO produces 2.5 m3 wastewater, so that 10 million tons CPO will produce 25 million m3 of wastewater. Palm Oil Mill Effluent (POME) has a highly organic character, e.g., COD of about 50,000 mg/L and BOD of about 25,000 mg/L. Thus, it can be concluded that POME is a serious threat to the environment and the quality of life in rural areas, unless proper pollution measures are taken. The current Indonesian effluent standard is 350 mg/L for COD and 100 mg/L for BOD5. Considering the high characteristic of POME, anaerobic process is the most suitable one (Metcalf and Eddy, 1991). Anaerobic process is a process where neither oxygen is present. A number of researchers studied POME treatment using different anaerobic methods. Conventional anaerobic process such as anaerobic ponds (Thanh, 1980; Southworth, 1979) and anaerobic digester could reduce BOD in the range of 80 to 95 % in hydraulic retention time (HRT) of 30 to 60 days. This lengthy retention time would need a very large digestion volume. Moreover, the application of anaerobic ponds has several problems, such as bad odors. In the conventional anaerobic digester processes, one of the most prevalent problems limiting the anaerobic digester (AD) performance is biomass washout. Improvements on the conventional reactor design generally involve some methods to selectively retain the solids in the digester. Membrane separation techniques afford an effective method to separate solids from the digester suspension and recycle them to the digester. So far several investigators have conducted experimental works of anaerobicmembrane process for treatment of a variety of wastewaters. From the results, application of membrane separation to anaerobic reactors was likely to be suitable for the treatment of the

2

wastewaters. The membrane bioreactor inherently allows no particulate matter to be expelled from the system. As a consequence, the particulate organic retained in the reactor can eventually be liquified and decomposed because of the long solid retention time (SRT). Also, the anaerobic microbes are able to proliferate without being washed out from the process. Despite the superior performance of membrane bioreactor, flux decline due to the concentration polarization and membrane fouling is the most important reason for the relatively slow acceptance of the systems. Especially, the filtration characteristics of the membrane bioreactor are more complicated owing to the role of various components presents in the AD broth (Choo and Lee, 1996). This disadvantageous is overcome by applying hydrodynamic technique such as increasing flow input in membrane and backflushing technique. To prevent using high energy and cost, backflushing technique is periodically developed. The newest backflushing technique with high intensity is known as rapid backflushing. This technique is developed for maintaining a stable flux during anaerobic treatment process for crude palm oil wastewater. Application of microfiltration membrane combining with backflushing technique in anaerobic membrane bioreactor system for industrial crude palm oil wastewater treatment gave the best opportunity to increase the performance of the system. Despite the advantages offered by the anaerobic membrane bioreactor (AMB), information on the effect of the biokinetic parameters for the system is lacking. Thus, the objective of this study was to study microfiltration membrane performance and determine the kinetic parameters.

MATERIALS AND METHODS Wastewater The wastewater was synthetic palm oil mill effluent which had COD and pH of 30,000


JURNAL ITENAS three-way valve, and timer. The apparatus configuration is shown in Figure 1. Influent was fed into the bioreactor through a peristaltic pump, then the mixed liquor from anaerobic bioreactor was flowed to a separator unit called microfiltration membrane through centrifugal pump. The pressure gauges were installed before and after the microfiltration membrane. The liquid which perforating membrane pore was called permeate, while the biomass solid which was retented and then flowed to bioreactor was called retentat. Backflushing technique was applicated in order to maintain membrane flux. Nitrogen for backflushing was brought outside by regulating the closing and opening timer in three-way valve. When the valve for permeate flow was closed, nitrogen flow would push the rest of permeate liquid into membrane side, and backflushing pressure would be shown in pressure measuring equipment in the cylinder.

mg/L and 7, respectively. The composition of the effluent was presented elsewhere (Ahmad, 2001).

Anaerobic Biomass Solid The biomass solid was obtained by developing biomass in anaerobic batch bioreactors to reach the biomass concentration about 3,000 mg MLSS/L.

Microfiltration Membrane The membrane was microfiltration membrane with hollow fiber configuration having 0.2 m diameter pore, 30 cm long 2 module, and 0.04 m total surface area. Microfiltration membrane material was polyetersulfone.

Experimental Apparatus Main apparatus for this research were bioreactors and membranes. Other apparatus were peristaltic pump, centrifugal pump,

Feed tank Aerator pump

Feed valve

Pressure gaugeValve Samplingl

Pressure gauge

Regulator solenoide valve

retentate Com

Membrane microfiltration

NC

Solenoide valve three way NO Permeate

Pressure gauge

Heater tank

Valve Pump

Figure 1 Anaerobic Membrane Bioreactor Configuration


3

JURNAL ITENAS

RESULTS AND DISCUSSION Change in Permeate Flux during Continuous Operation Figure 2 shows the changes of permeate flux over the 17 days of operation. The flux 2 were attained in the range of 12 to 16 L/m .hr depended on SRT. Increasing of SRT will enhance the MLSS concentration. The permeate flux decreased as the MLSS concentration increased. This phenomenon was called concentration polarization causing by insolubilization of high molecular substrate. The increase of MLSS enhanced the deposition of cells into gel layer matrix and accelerated the deterioration of flux. This result was agree with that reported by Li et. al. (1985) who found a proportional decrease in membrane flux with an increasing in digester suspended solids. Other factors which enhanced the decreased of flux was the lower pH. Decreasing pH under 7 accelerated fouling because it will be followed by the increasing adsorption.

4

28 24 Flux (L/m2.hr)

Operating Procedures The research methods included experiments with batch and continuous process using the anaerobic membrane system. The batch process was conducted in a tank with working volume of 10 L to obtain biomass concentration up to 3,000 mg MLSS/L. The continuous process was conducted in anaerob tanks with working volume of 10 L and HRT of 1 and 2.5 days. The SRT was varied from 25, 35, 45, and 55 days. The membrane used for microfiltration process was hollow fiber made from polyetersulfone. Backflushing technique was applied with giving pressured air from permeate side of the system. The optimum condition of membrane process was obtained in trans membrane pressure (TMP) of 0.4 bar. The backflushing optimum condition was obtained at backflush pressure, time, and interval of 1.6 bar, 1 second, and 2 minutes, respectively (Sumihar et al., 2001).

20 16 12 8 4

SRT 25 days SRT 35 days SRT 45 days SRT 55days

0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 Operating time (days)

Figure 2 Profile Of Flux Against Operating Time

Reactor Performance Figure 3 shows bioreactor performance during the process. The concentration of MLSS in reactor increased between 2,000 and 25,000 mg/L as the increasing of SRT and HRT. HRT of 2.5 days gave lower COD removal than that of HRT of l day, because the lower HRT would gave the lower substrate. MLSS in 1 day HRT could have higher concentration than that of 2.5 day HRT. Beside that, the higher SRT would also give the higher MLSS. This indicated that higher SRT would give lower sludge disposal. Therefore, the MLSS would be higher at the higher SRT. On the other hand, the obtained percentage of COD removal showed that the increasing HRT did not influence the percentage of COD removal. The COD removal percentage was between 89 and 91 percent in HRT between 1 and 2.5 days. Gas Content Figure 4 shows methane content in biogas, which was ranged from 5 to 14 percent. This composition was lower than that reported by other workers (Borja-Padilla and Banks, 1993; Ng et. al., 1985), who investigated crude palm oil effluent in longer HRTs (6 to 40 days). Another factor that caused lower gas content was hydrogen pressure. The hydrogen pressure in this research was more than 10-4 atm, which decreased acetate formation and then methane production, because acetate is a substrate source for methane formation.


JURNAL ITENAS

40

35 30

30

25 20

20

15 10

MLSS (g/L)

g/day COD removal (x 10g/days)

40

10

5 0

0 20

25

30

35

40

45

50

55

60

Solid Retention Time (days) COD removal at HRT1days CODremoval at HRT 2,5 days MLSS at HRT 2,5 days MLSS at HRT 1 days

100 95 90 85 80 75 70 65 60 55 50

45 40

% COD removal

CODremoval (x 10 g/days)

50

35 30 25 20 15 10 20

25

30

35

40

45

50

55

60

Solid Retention Time (days) g/days COD removal at HRT 1day g/day COD removal at HRT2,5 days % COD removal at HRT 1days % COD removal; HRT 2,5 days

Figure 3 Performance Of Bioreactor Against Solid Retention Time

45 40

HRT 1 days

% Metana

35

Reduction in Bacterial Size Figure 5 shows the reduction of bacterial size. Before operating with membrane, the bacterial size was 36 50 m and the distance among two flocs was 2 3 m. After the operation, the size was reduced to 6 11 m and the distance was 6 11 m. Decreasing bacterial size was caused by shear stress and circulation in the pumping unit. The similar result was found by Ross (1992) having particle size reduction up to 55 % after biomass passed the membrane. General Discussion This result is different from Razi et. al. (1999), who investigated POME by using anaerobic membrane bioreactor. The differences are in COD removal, methane composition, and MLSS concentration. By using anaerobic membrane bioreactor, Razi could give COD removal up to 93%, methane composition up to 74.5%, and MLSS up to 56,600 mg/L. This research gave COD removal, methane composition, and MLSS up to 89%, 14%, and 28,000 mg/L, respectively. Different value between Razi's and this research was caused by several factors, such as higher HRT and lower SRT using in this research. Higher HRT made higher organic loading rate than those of Razi's experiments. Lower HRT would provide not enough time for subsrate degradation and the hydrogen production would increase, which decreased methane production. Lower SRT increased sludge wasting rate, which would decrease the MLSS concentration in bioreactor.

HRT 2.5 days

30

Applying material balances to anaerobic membrane bioreactor with cell recyle as shown in Figure 6, the following biokinetic model for completely mixed system was developed :

25 20 15 10 5 0 20

25

30

35

40

45

50

55

60

Solid Retention Time (days)

Figure 4 Production Of Methane Against Solid Retention Time


Material balances in anaerobic bioreactor : Accumulation = input feed + input resirculation - output output bioreactor + reaction rate

5

JURNAL ITENAS And by substituting equation 3 and 9 to equation ,

m=

m m .S Ks + S

,then we will get:

1 + kd ) qc S= 1 m m - ( + kd ) qc Ks (

(a)

(10)

The biomass concentration in the bioreactor was derived from the equation (Chin, 1981):

X =

q cY ( So - Se ) q (1 + kdq c )

(11) Fr, Sr, Xvr

Fe,Xe membrane Fo, So, Xo (b)

V, S, X

F – Fw + Fr

Figure 5 Anaerobic Bacterial Floc (a) After 7 Days; (b) After 18 Days (Steady State

Anaerobic bioreactor

(1) Fr.Xvr - (F-Fw-Fr) Xv - FwXv + mXv.V - b Xv.V- kd Xv.V = 0 (2)

Fw

substitute = Fr/F and = V/F to equation 2, then equation 2 will be : m= G(1/) + kd G = 1 + - (Xvr/Xv)

(3) (4)

Material balances in membrane, with assume that cell mass completely retained: Xe = 0 (F-Fw + F).Xv - FXvr = 0 1+ a- a(Xvr/Xv) = Fw/F G = Fw/F

(5) (6) (7)

The sludge wastage rate conducted directly from the bioreactor: Oc = V.Xv / Fw.Xv = V/Fw G =t /qc

6

(8) (9)

Figure 6 Material Balances Of Anaerobic Membrane Bioreactor

In this: Fo = Volumetric influent flow rate; Fr = Sludge recirculation volumetric flow rate; Fw = Sludge disposal volumetric flow rate; X = Biomass concentration in bioreactor; Xo = Biomass concentration in feed; V = Bioreactor volume; S = Substrate concentration; Fe = Effluent flow rate; Xe = Biomass concentration in effluent. The kinetic parameters consist of specific growth rate (mm), yield (Y), half saturated constant (Ks), and maximum substrate utilizing (k). Yield and specific decay rate were determined by plotting (So-S)/Xq) as a function of (1/qc). Yield can be calculated from the slope of the straight line and the intercept will give the specific decay rate value as ilustrated in Figure 7 and 8. Specific growth


JURNAL ITENAS substrate source could be explained from the ratio of BOD/COD. The ratio of BOD/COD in the effluent is 0.38. Substrate with BOD/COD ratio higher than 0.65 is classified as a readily biodegradable compound, so this susbtrate is a slowly degradable compound that will has high Ks. The lower mm value might be influenced by the shorter sludge retention time. The shorter SRT will increase the higher sludge wastage which caused decreasing of growth rate.

S/(1/WTC + kd) (mgCOD/L.hari-1)

200000 150000 100000 y = 76.722x - 156904 R2 = 0.7902

50000 0 0

1000

2000

3000

4000

5000

S (mgCOD/L)

Figure 7 Y and kd at HRT of 1 Day

1.4 (So-Se)/X.HRT) (gCOD/gMLSS.days)

rate (mm) and half saturated constant (Ks) were determined by plotting (S/(1/c+kd) as a function of S. mm can be calculated from the slope of straight line and the intercept will give the Ks/mm. k was the divided between m m/Y as ilustrated in Figure 9 and 10 and Table 2 also present the value of kinetic coefficient for an anaerobic with different process. Table 1 presents the values of Y, kd, Ks, and m max for an o AMB treating syntetic POME at 35 C. The Ks value obtained in this investigation are about 2.52 and 2.05 g/L. This value was not so difference with Chin (1981) who obtained 4,032 g/L but higher than that of Boopathy and Tilche (1992) and Shieh et.al. (1985). The difference in Ks value was effected by the different substrate source and concentration of substrate. Higher concentration of substrate will increase the value of Ks. The Specific growth rate value attained in this research is about 2,25.10-2 and 1,3.10-2 day. This value are lower than those of Boopathy and Tilche (1992), Shieh et. al. (1985). The difference in mm value was also effected by the different substrate source and substrate concentration. Higher concentration of substrate will increase the value of m. In this research, mm value was very low while Ks value was very high. The Ks value is the saturation constant, equal to substrate concentration when the rate of substrate utilization is one half of the maximum rate. m and Ks value were depend on the substrate source and substrate concentration. Microorganism with slow growing organism will have small mm and high Ks. Palm oil effluent has a high concentration of substrate, so it will give high Ks value. The effect of

1.2 1 0.8 0.6 0.4 y = 25.478x + 0.1485 R2 = 0.7514

0.2 0 0

0.01

0.02

0.03

0.04

0.05

-1

1/SRT (days )

Figure 8 Y and kd at HRT of 2.5 Days

Table 1 Kinetic Parameters Coefficient For The Anaerobic Membrane Bioreactor (AMB) HRT Y (mgCOD/mgMLSS kd (days-1) Ks (g/L) mmaks (days-1)


7

100000 80000 60000

y = 47.354x - 121466 2 R = 0.6078

40000 20000 0 4000

4200

4400

4600

4800

200000

-1

S/(1/WTC + kd)

-1

S/(1/WTC + kd)

(mgCOD/(L.hari )

120000

(mgCOD/L.days )

JURNAL ITENAS

150000 100000 50000 y = 76.722x - 156904

R2 = 0.7902 0

5000

0

1000

S (mgCOD/L)

2000

3000

4000

5000

S (mgCOD/L)

Figure 9 Ks and mmaks at HRT of 1 Day

Figure 10 Ks and mmaks at HRT of 2.5 Days

Table 2 Anaerobic Bioreactor Kinetic Coefficient

Bioreactor Type

-

Y GVSS/gCOD 0.035

kd days-1 0.027

Chin, 1981

4.032

-

0.140

0.037

Chin, 1981

Synthetic Molase

0.154 0.383

0.16 0.296

0.08 1.227

0.09 0.253

UASB UASB

Sugar Plant Etanol Plant

48.9 -

3.121 -

0.191

0.009

Baffle

Oil and Fat

1.060

0.187

0.395

0.027

Shieh, 1985 Boopathy and Tilche, 1992 Riera et. al., 1985 Callander et. al., 1987 Adrianto et. al., 2001 This research

Digester with cell recirculation Digester without cell recirculation Fluidized Bed HABR

Type of wastewater Palm oil

KS g/L 9.65

Palm oil

mm days-1

Syntetic (HRT 2.52 2.25 .10-3 0.0012 0.02 of 1 day) AMB Sintetik (HRT 2.05 1.3 .10-3 0.0392 0.005 of 2.5 day) HABR = hybrid anaerobic baffled reactor; UASB= upflow anaerobic sludge blanket AMB

CONCLUSIONS 1. The flux were attained in the range of 12 to 16 L/m2.hr and the anaerobic bioreactor membrane was able to reduce COD up to 89% for several SRT. This process showed that decreasing of HRT would not influence the bioreactor performance. 2. Methane was produced very low in this process, which was caused by higher hydrogen pressure in bioreactor.

8

References

This research

3. During the process, bacterial floc size was reduced from 36 50 m to 6 - 11 m. 4. Kinetic parameters depended on the HRT. At HRT of 1 day, the maximum specific growth (mm), yield (Y), specific decay rate (kd) and constant half saturation (Ks) were -2 -1 -3 2.25 .10 day , 1.22 .10 mg COD/mg MLSS, 2.03 .10-2 day-1 and 2.52 .10-3 mg/L, respectively. At HRT of 2.5 days, mm, Y, kd and Ks were 1.3 .10-2 day-1, 3.92 .10-3 mg


JURNAL ITENAS -2

-1

3

COD/mg MLSS, 5.83 .10 day and 2.05 .10 mg/L, respectively.

ACKNOWLEDGMENT This project was funded by Indonesian Government through RUT VIII 2001-2003 with contract no. 011.11/SK/RUT/2001.

REFERENCES Ahmad, A. 2001. Biodegradasi Limbah Cair Industri Minyak Sawit Dalam Sistem Bioreaktor Anaerob. Disertasi, Program Pascasarjana, Institut Teknologi Bandung. Boopathy, R. dan Tilche, A. 1992. Pelletization of biomass hybrid anaerobic baffled reactor (HABR) treating acidified wastewater. Bioresource Technol., 40(2), 101-107. Borja-Padilla, R. and Banks, C. J. 1993. Thermophilic Semi-continuous Anaerobic Treatment of POME. Biotechnology Letters, 15 (7), 761-766. Callander, I.J., Clark, T.A., dan McFarlane, P.N. 1987. Anaerobic digestion of wood ethanol stillage using upflow anaerobic sludge blanket reactor. Biotechnol. Bioeng., 30, 896-908. Chin, K.K. 1981. Anaerobic treatment kinetics of palm oil sludge, Wat. Res., 15, 199-202. Choo,K.H. and Lee, C.H. 1996. Effect of anaerobic digestion broth composition on membrane permeability, Wat. Sci. Tech., 34, 173-179. Li, A., Kothari, D., and Corrado, J.J. 1985. Application of membrane anaerobic reactor system for the treating of -


industrial wastewaters. Proceedings of 39th Purdue Industrial Waste Conference, Lafayette Indiana, May. 627 636. Metcalf and Eddy, Inc. 1991 Wastewater Enggineering Treatment, Disposal, rd Reuse. 3 ed. NewYork: McGraw-Hill, Inc. Ng, W.J., Wong, K.K. and Chin, K.K. 1985. Two phase anaerobic treatment kinetics of palm oil wastes. Wat. Sci. Tech., 19(5),667-669. Razi, A. F. and Noor, M.J.M.M. 1999. Treatment of Palm Oil Mill Effluent (POME) with The Membrane Anaerobic System. Wat. Sci. Tech. Vol. 39, No. 10-11, hal. 159-163. Ross, W.R., Barnard, J.P., Strohwald, K., Grobler, C.J. and Sanetra, J. 1992. Practical application of the ADUF process to the full-scale treatment of a maize-processing effluent. Wat. Sci. Tech., 25(10),329-337. Shieh, W.K., Li, C.T. dan Chen, S.J. 1985. Performance evaluation of the anaerobic fluidised bed system: III. process kinetics. J. Chem. Tech. Biotechnol., 35B, 229-234. Southworth, A. 1979. Palm Oil Factory Effluent Treatment by Anaerobic Digestion in Lagoons, Proceeding 35 th Ind.Waste Conf., Purdue University, Purdue. Sumihar, H.D.S., Ahmad, A., Setiadi, T. and Wenten, I.G. 2001. Treatment of Palm Oil Effluent by Anaerobic Membrane Bioreactor. Proceeding Regional Symposium Chemical Engineering, BP 11-1. Thanh, N.C. 1980. High Organic Wastewater Control and Management in the Tropics. Water Pollution Control, CDG, AIT-ERL, Bangkok, November.

9