Environ. Sci. Technol. 2010, 44, 335–342
Operational Boundaries for Nitrite Accumulation in Nitrification Based on Minimum/Maximum Substrate Concentrations That Include Effects of Oxygen Limitation, pH, and Free Ammonia and Free Nitrous Acid Inhibition S E O N G J U N P A R K , † W O O K E U N B A E , * ,‡ AND BRUCE E. RITTMANN† Center for Environmental Biotechnology, Biodesign Institute at Arizona State University, 1001 South McAllister Avenue, Tempe, Arizona 85287-5701, and Department of Civil & Environmental Engineering, Hanyang University, Sa 1-Dong, Ansan, Gyunggi-Do, Republic of Korea
Received August 8, 2009. Revised manuscript received November 8, 2009. Accepted November 11, 2009.
Recent studies on shortcut biological nitrogen removal (SBNR), which use the concept of denitrification from nitrite, have reported the key factors affecting nitrite build-up, such as dissolved oxygen (DO) limitation, pH, and free ammonia (FA) and free nitrous acid (FNA) inhibition. This study extends the concept of the traditional minimum substrate concentration (Smin) to explain the simultaneous effect of those factors. Thus, we introduce the minimum DO concentration (DOmin) and the maximum substrate concentration (Smax) that are needed to support a steady-state biological system. We define all three values as the MSC values. The model provides a method to identify good combinations of pH, DO, and total ammonium nitrogen (TAN) to support shortcut nitritation. We use MSC curves to show that the effect of DO-alone and the effect of DO plus direct pH inhibition cannot give strong enough selection against nitrite oxidizing bacteria to work in a practical setting. However, adding the FA and FNA effects gives a strong selection effect that is accentuated near pH 8. Thus, a generalized conclusion is that having pH ∼8 is favorable in many situations. We defined a specific operational boundary to achieve shortcut nitritation coupled to anaerobic ammonium oxidation (ANAMMOX), in which the effluent concentrations of total nitrite and total ammonium should be approximately equal. Experimental results for alkaline and acidic nitrite-accumulating systems match the trends from the MSC approach. In particular, acidic systems had to maintain higher total ammonium, total nitrite, and DO concentrations. The MSC values are a practical tool to define the operational boundaries for selecting ammonium-oxidizing bacteria while suppressing nitrite-oxidizing bacteria.
* Corresponding author e-mail:
[email protected]; phone: +82-31-400-5148; fax: +82-31-417-8139. † Biodesign Institute at Arizona State University. ‡ Hanyang University. 10.1021/es9024244
2010 American Chemical Society
Published on Web 12/08/2009
Introduction Dual limitation with electron donor and electron acceptor slows the reaction rate and is common in engineered nitrification systems. This situation is further complicated because the nitrifying bacteria are divided into two distinct groups, ammonium-oxidizing bacteria (AOB) and nitriteoxidizing bacteria (NOB), that have a relationship which is synergistic and competitive. The AOB utilize ammonium (NH4+) as their donor and release nitrite (NO2-) as the oxidized product. The NOB utilize nitrite as their donor, with nitrate (NO3-) being the oxidized product. AOB and NOB utilize dissolved oxygen (O2) as the electron acceptor. Thus, the NOB must have the AOB to supply their donor substrate, but the AOB and NOB compete for the common acceptor. In the competition for O2, it appears that the NOB often are at a disadvantage. For example, Hanaki et al. (1) measured the substrate utilization rate of nitrifiers cultivated in a suspended-growth system with a 0.5-mg/L DO concentration. The NOB were more severely limited, resulting in nitrite accumulation. A similar phenomenon was observed in a biofilm reactor by Bernet et al. (2), who reported that, when the DO concentration was elevated from 0.5 to 5 mg/L, nitrite accumulation was relaxed so that nitrate was produced. When the DO concentration was dropped again to 0.5 mg/L, nitrite accumulation resumed. Table 1 summarizes different DO concentrations suggested for nitrite accumulation; the values range from 0.4 to 5 mgDO/L. In some situations, suppressing NOB so that nitrite is not oxidized to nitrate is a desired outcome. The common example is shortcut nitritation/denitrification (also known as shortcut biological nitrogen removal, SBNR), in which the nitrite generated by AOB is reduced to N2 gas by denitrifying bacteria (10, 13). The goals of shortcut approaches are to minimize the requirements to supply O2 for nitrification and an organic donor to drive denitrification. Recently, controlling the DO concentration as a means to affect shortcut nitrification/denitrification was proposed and demonstrated (10, 14, 15). DO control augments the more traditional controls using pH, temperature, free ammonia (FA) concentration, and free nitrous acid (FNA) concentration to suppress NOB and build up nitrite (10, 13, 16-18). While the principles of NOB suppression are known, we do not yet have good definition of what DO, FA, and FNA concentrations must be maintained to suppress NOB without also harming the AOB. An important concept for understanding when a microbial type is sustained or suppressed is the minimum electrondonor substrate concentration to support steady-state biomass (Smin) (19, 20). A substrate concentration smaller than Smin causes the microorganisms to have a negative growth rate, which means that they are gradually washed out. In this study, we expand the concept of Smin to include oxygen limitation and inhibition by pH, FA, and FNA. We identify the required range of DO concentration by using simulations with reported kinetic parameters of nitrifiers, and compare the predicted range with experimental results from the literature.
Modeling Approach Expanding the Concept of Minimum Substrate Concentration with Oxygen Limitation. Under single-limitation, Smin can be derived from the Monod equation as eq 1a for the donor substrate or eq 1b for the acceptor substrate (20). SD,min ) KED ·
b Yqˆ - b
VOL. 44, NO. 1, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
(1a) 9
335
SA,min ) KEA ·
b Yqˆ - b
(1b)
in which SD,min or SA,min is the minimum substrate of electron donor and electron acceptor, respectively. KED and KEA are half-maximum rate concentrations for the electron donor (SED) and electron acceptor (SEA, the oxygen concentration in aerobic condition), respectively. Y is the true yield coefficient, b is the endogenous decay rate, and qˆ is the maximum specific substrate utilization rate. Under dual limitation by electron-donor and electronacceptor substrates, the specific electron-donor utilization rate (q) is expressed with a multiplicative Monod model (21). q ) qˆ
qAOB )
SEA SED (KED + SED) (KEA + SEA)
(2)
In eq 2, q is controlled by the concentration of electron acceptor, resulting in an apparent lowering of the maximum specific substrate utilization rate (qˆ) to qˆ[SEA/(KEA + SEA)]. On the basis of the apparent value of qˆ, minimum substrate concentrations are expressed as eqs 3a and 3b under dual limitation. The detailed derivation of eqs 3a and 3b is shown in Supporting Information. SD,min
b ) KED · qˆ · SEA Y· -b (KEA + SEA)
SA,min ) KEA ·
(3b)
From eqs 3a and 3b, we see that the minimum electron donor-substrate concentration is affected by the concentration of the acceptor and vice versa. Direct Inhibition and Indirect FA and FNA Inhibition with pH. The pH can affect nitrification reactions in two ways: (1) directly by changing the enzyme reaction mechanism (17, 22-24) and (2) indirectly by changing the concentrations of the inhibiting species, FA and FNA (13, 17, 25, 26). We represent the direct effect, which has had comprehensive research (17, 22-24), by allowing the maximum specific substrate utilization rate of AOB or NOB to be affected by pH in a way that can be captured by the empirical bellshaped equation of eq 4 (24) qˆobs
{
[
( (
KS,AOB 1 +
)
qˆobs,AOBSNHX
(
))
IFNA IFNA IFA + SNHX 1 + + KI,FNA,AOB KI,FNA,AOB KI,FA,AOB (5a)
qNOB )
( (
KS,NOB 1 +
IFNA KI,FNA,NOB
)
qˆobs,NOBSNO2
(
+ SNO2 1 +
IFNA KI,FNA,NOB
(3a)
b qˆ · SED Y· -b (KED + SED)
qˆmax π ) 1 + cos × (pH - pHopt) 2 w
in which qˆobs and qˆmax are, respectively, the maximum specific substrate utilization rate at a given pH and for the optimal pH, and w is the pH range within which the qˆobs is larger than one-half of qˆmax. To calculate Smin, qˆobs in eq 4 is substituted for qˆ in eqs 3a or 3b. Park et al. (24) used eq 4 to describe experimental data from 7 different cultivation systems, and they estimated average parameter values that gave good simulations of the experimental results. The best way to represent FA and FNA inhibition is not completely resolved. Recently, Park and Bae (26) proposed the simultaneous-effect model, shown in eqs 5a and 5b, for how the FA and FNA concentrations affect the maximum specific utilization rates for AOB and NOB, respectively:
]}
+
IFA
))
KI,FA,NOB (5b)
in which, q and qˆ, respectively, are the specific and the maximum specific substrate utilization rate (mgN/mgVSSd), S is a substrate concentration, KS is the half-maximumrate concentration for the substrate, and KI is the inhibition concentration for the inhibitor. IFA or IFNA is the FA or FNA concentration (mgFA/L or mgFNA/L), and it is clearly identified as the inhibitor. qˆobs is the maximum specific substrate utilization rate at a given pH in eq 4. Since FA and FNA concentration are affected by pH, these models also represent the indirect pH effect by calculating FA and FNA concentrations according to the pH (26). Integration of Effects. Since AOB and NOB utilize the same electron acceptor, the form of the SA,min equation is the same for both, as shown in eq 6a, for which the subscript i refers to either AOB or NOB. SA,min ) Yi ·
[ (
KED,i · 1 +
)
bi · KEA qˆobs,i · SED
(
IFNA IFNA IFA + SED · 1 + + KI,FNA,i KI,FNA,i KI,FA,i
(4)
)]
- bi (6a)
Smin for each electron donor-substrate can be calculated from eqs 6b and 6c for AOB and NOB, respectively. The detailed
(pHopt - w e pH e pHopt + w or |pH - pHopt | < w)
TABLE 1. DO Concentrations Causing Nitrite Accumulation reference
DO (mg/L)
effect*
Hanaki et al. (1) Ruiz et al. (3) Ciudad et al. (4) Blackburne et al. (5) Kim et al. (6) Garrido et al. (7) Joo et al. (8) Bernet et al. (2) Fux et al. (9) Chung et al. (10) Gali et al. (11) Yamamoto et al. (12)
0.5 0.7 1.4 0.4 1.0 1.5 2.0-5.0 0.5 2.0-4.0 3.0 5.0
inhibition of nitrite oxidation and its accumulation 93% as NO2-, 67% of NH4+ 75% as NO2-, 95% of NH4+ 15%-95% as NO2100% as NO2100% as NO2-, 50% of NH4+ 100% as NO2-, 60% of NH4+ 90% as NO2-, 100% of NH4+ 100% as NO2-, 50% of NH4+ 93% as NO2-, 88% of NH4+ 100% as NO2-, 50% of NH4+ 93%% as NO2-, 43% of NH4+
* Note: Accumulation rate of nitrite (NO2-/NOx-), and removal rate of ammonia.
336
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 1, 2010
system suspended growth activated sludge biofilm airlift reactor biological aerated filter completely stirred biofilm reactor moving bed biofilm reactor activated sludge with biofilm carriers sequential batch reactor up-flow reactor with biomass carrier
TABLE 2. Conditions for Each Simulation Step case
electron donora
pH
I
not applicable
II III
7-9 (interval: 1)
inhibition
TAN or TNN: 0-1000 mgN/L(interval: 1 mgN/L)
not applicable
species
eq no.
AOB and NOB
2 and 3
not applicable FAb: 0-449 mg/L FNAb: 0-0.73 mg/L
a
TAN: Total ammonium nitrogen. TNN: Total nitrite nitrogen. 7a or 7b, respectively.
b
2-4 5 and 6
Entire range calculated by pH and TAN or TNN from eqs
TABLE 3. Kinetic Parameters for the Model Simulations kinetic parameters Y qˆmax KS KDO b w pHopt KI,FA KI,FNA
yield coefficient (mgVSS/mgN-d) maximum specific substrate utilization rate at optimal pH (mgN/mgVSS-d) half-max-rate concentration (mgN/L) half-max-rate concentration (mgDO/L) decay coefficient (d-1) pH range optimal pH inhibition concentration (mgFA/L) inhibition concentration (mgFNA/L)
derivations are shown in the Supporting Information. Because of the quadratic formula of eqs 6b and 6c, each electron donor-substrate concentration has two limiting values with FA and FNA inhibition. The lower limiting value (SD,min) retains the same meaning as the traditional Smin: the minimum substrate concentration able to achieve steady-state biomass in a continuous stirred tank reactor. However, SD,max is the maximum substrate concentration able to sustain steadystate biomass when FA and FNA inhibition are important.
ZAOB2 - 4 · KED ·
-ZAOB SD,min )
(
b2 · 1 +
IFNA KI,FNA,AOB
fFA(pH) 2· ·b KI,FA,AOB
ZAOB2 - 4 · KED ·
-ZAOB + SD,max )
fFA(pH) · KI,FA,AOB
(
b2 · 1 +
2·
fFA(pH) ·b KI,FA,AOB
)
fFA(pH) · KI,FA,AOB IFNA KI,FA,AOB
)
for AOB
(6b) -ZNOB SD,min )
ZNOB2 - 4 · KED · 2·
-ZNOB + SD,max )
fFNA(pH) ·b KI,FNA,NOB
ZNOB2 - 4 · KED · 2·
fFNA(pH) · b2 KI,FNA,NOB
fFNA(pH) ·b KI,FNA,NOB
fFNA(pH) · b2 KI,FNA,NOB
AOB
NOB
0.33
0.083
3.1 1.5 0.5 0.15 3.2 8.4 10 0.5
refs
13 2.7 0.68 0.15 2.4 7.7 0.75 0.1
Rittmann and McCarty (20)
Park et al. (24) Hellinga et al. (13), Van Hulle et al. (17), Carrera et al. (25), Park and Bae (26)
fFA(pH) ) (17/14) · [10pH/(exp (6334/(273 + °C)) + 10pH], and fFNA(pH) ) (47/14) · {1/([exp (-2300/(273 + °C)) × 10pH] + 1}. Finally, the traditional concept of minimum substrate concentration produces three different minimum/maximum substrate concentrations (the MSC values).
Modeling Experiments Modeling Tests for the MSC Values. Equations 6a-6c show that the MSC values are affected by the parameters values. The MSC values for a microbial strain will be unique because each strain has its own unique parameter values. To characterize how the minimum substrate concentration affects AOB versus NOB, we carried out modeling experiments in three cases: I. Including only dual limitation by dissolved oxygen (eq 3). II. Adding the direct affect of pH (eq 4 into eq 3a). III. Adding FA and FNA inhibition (eq 6a). In case I, we used two MSC values (SD,min and SA,min) to describe the interaction between AOB and NOB. The direct pH effect is included in case II, and FA and FNA inhibition effects are additionally included in case III. Table 2 shows the simulation conditions for each case, and all simulations were coded and computed in Matlab 7.0. For cases I-III, the electron-donor substrate was varied from 0 to 1000 mgN/L for AOB and NOB. For cases II and III, we altered the maximum specific substrate utilization rate by using pH 7, 8, or 9. For case III, FA and FNA concentrations were calculated with eqs 7a and 7b, respectively, on the basis of the pH value and the total ammonium nitrogen (TAN) or total nitrite nitrogen (TNN) concentration (26). FA )
for NOB
FNA )
TAN × 10pH 17 14 [exp(6334/(273 + °C)) + 10pH]
(7a)
TNN 47 14 [exp(-2300/(273 + °C)) × 10pH] + 1
(6c)
(7b)
in which, i ) AOB or NOB, ZAOB ) b · [1 + (IFNA/KI,FNA,AOB)] Y · qˆobs,AOB · (SEA)/(KEA + SEA), ZNOB ) b · [1 +(IFA/KI,FA,NOB) + (fFNA(pH) · KED)/(KI,FNA,NOB)] - Y · qˆobs,NOB · (SEA)/(KEA + SEA),
The reported kinetic parameters we used are listed in Table 3. For w and pHopt of the pH model, the values are the averages for the three different mixed cultures in Park et al. VOL. 44, NO. 1, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
337
FIGURE 1. Minimum substrate concentration curve for AOB under dual limitation by donor (TAN) and acceptor (O2) only (Y ) 0.33 mg VSS/mg S-d; qˆmax ) 3.1 mg S/mg VSS-d; KS ) 1.5 mg/ L; KDO ) 0.5 mgDO/L; b ) 0.15 d-1). Washout occurs for any combination of donor and acceptor concentration lying below the curve. The dotted straight lines show the limiting values for single limitation by either substrate.
FIGURE 2. Comparison of the minimum substrate concentration curves of AOB and NOB with DO limitation. Region I: AOB and NOB wash out. Region II: AOB survives, but NOB washes out. Region III: Both survive.
(24). The values for the inhibition concentrations are the averages among the references listed.
Results and Discussion MSC Values under Dual Limitation Alone: Case I. With only dual limitation by DO (eqs 3a and 3b), the MSC values for one substrate depends on the concentration of the other substrate. This is shown by the “minimum substrate concentration curve” in Figure 1. For example, when the DO is 1 mg/L, the SD,min value is 0.43 mg TAN/L, but SD,min increases to 1.0 mg TAN/L when DO is 0.25 mg/L. However, if the electron-donor or electron-acceptor concentration is high enough, the Smin of the other substrate approaches the traditional Smin concentration, which is shown by the dotted straight lines in the figure. Any combination of donor and acceptor concentrations that falls below the MSC curve indicates system failure or washout of the microorganism. For example, if system is maintained at 0.43 mg N/L of TAN, the DO should be maintained above 1 mg O2/L to avoid washout of AOB. This DO concentration is much higher than the DOmin concentration (0.09 mg O2/L) that is calculated by eq 1a for single limitation. Figure 2 compares the MSC values for AOB and NOB. If the donor and DO concentrations are maintained above both curves (noted as region III), AOB and NOB can coexist. Both microorganisms wash out if the concentrations are below both curves (region I). However, if the concentrations lie between the two curves, noted as region II, the AOB survive, but the NOB washout, and the system builds up nitrite. This points out the potential of using the DO concentration to create shortcut nitritation, which avoids NO3-. For instance, if the TAN and TNN concentrations are the same at 0.75 mgN/L, a DO concentration of 1 mgO2/L leads to shortcut nitritation, but 2 mgO2/L leads to traditional two-step nitrification to NO3-. However, region II is narrow in Figure 2, which makes it difficult to achieve shortcut nitritation reliably with DO control alone in a real-world treatment plant. pH Change Affecting the MSC Curve: Case II. By substituting eq 4 into eq 3a, the MSC curve includes the direct effect of pH inhibition. Figure 3 shows the MSC curves for AOB and NOB with pH direct inhibition. For AOB (Figure 338
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 1, 2010
FIGURE 3. Change of MSC curve with pH for AOB (a) and NOB (b). The inset figure represents the MSC curve for less than 10 mg/L concentration.
3a), DOmin is higher at pH 7 than at pH 8 and pH 9. This result occurs because 8.4 is the optimal pH, pHopt, for AOB (24). On the other hand, the DOmin for NOB is highest at pH 9, since its pHopt is 7.7 (24). The comparison in Figure 3 suggests that selection for AOB and against NOB can be made by keeping
FIGURE 4. Comparison with MSC curves at different pHs (a, pH 7; b, pH 8; c, pH 9) when TNN or TAN is fixed at 50 mg/L and the other nitrogen species has variable concentration from 0 to 1000 mg N/L. The temperature is 25 °C. DOmin for NOB is near infinity at pH 9 because of FA inhibition. the pH near 9 and the DO around 0.5 mg/L, if TAN and TNN are near 5 mgN/L. Although pH 9 is favorable for shortcut nitritation, the region of selective growth for AOB is still narrow in Figure 3. For example, if TAN and TNN are maintained at 50 mg/L, Figure 3a and b gives DOmin of 0.10 mg/L for AOB and 0.35 mg/L for NOB, respectively. Thus, the DO window for successful shortcut nitritation is narrow and probably not attainable in a practical setting. The situation is somewhat better for low total-N concentration. For example, DOmin values are 0.75 mg/L for NOB and 0.10 mg/L for AOB with each substrate-nitrogen of 5 mg/L. However, this situation is unlikely to be sustainable in a practical setting. MSC Curve for Simultaneous Effects of pH, Oxygen, and FA and FNA Inhibition: Case III. As shown in Figure 4, adding inhibition by FA and FNA shows the most dramatic differences between the MSC curves of AOB and NOB. The differences become dramatic because (from eqs 7a and 7b) FA inhibition increases with pH, but FNA inhibition increases as pH is decreased. As discussed for eq 6c, Figure 4 shows a maximum substrate concentration, SD,max, that represents
the boundary for supporting the biomass in a system affected by inhibition. SD,min continues to exist, although it is barely seen because of its low value. If the TAN or TNN concentration is below SD,min or over SD,max, DOmin approaches infinity, which identifies the washout range, WR. For TAN or TNN less than 10 mg/L, the MSC for DO is the same as with Figure 3. The left side in Figure 4 simulates the effect for AOB and NOB in an ammonia-rich system, that is, when TAN is allowed to increase to as high as 1000 mg/L, while TNN is fixed at 50 mg/L. The right side simulates the effect in a nitrite-rich system, that is, when TNN can go as high as 1000 mg/L, while TAN is fixed at 50 mg/L. Figure 4 shows that the ammonia-rich condition gives more severe inhibition with pH and TAN concentration for both nitrifiers than does the nitrite-rich condition. In both cases, pH 9 causes washout of NOB. In case of an ammonia-rich condition, when DO is 2 mg/ L, pH 7 gives shortcut nitritation only for TAN greater than about 300 mg/L. The TAN concentration for shortcut nitritation is lowered to about 25 m/L for pH 8, while the AOB have a SD,max value of over 600 mg/L. pH 9 is not VOL. 44, NO. 1, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
339
FIGURE 5. Contour of MSC curve between TAN and TNN concentration at the generally optimal pH of 8. The temperature is 25 °C. The values on the line are DOmin for AOB or NOB, as noted. suitable for shortcut nitritation, due to the AOB washout with a TAN concentration above 100 mg/L. Therefore, pH 8 is favorable to operate a shortcut nitritation system. For two years, Chung et al. (10) successfully operated a SBNR process that used shortcut nitritation with a 2 mg/L DO concentration, around 250 mg N/L TAN, pH 8, and a TNN concentration around 170 mg/L. For the conditions of Chung et al. (10), the model compute a DOmin values of 0.52 for AOB and WR for NOB, which agree with the experimental results. In the case of a nitrite-rich system, again the NOB are always preferentially washed out, since FNA inhibition for NOB is severer than that for AOB. Although it should be possible to achieve shortcut nitritation at any pH with a nitrite-rich condition, the optimal pH is around 8, since the DOmin concentration for the AOB is very low and affected only a little by the TNN concentration. On the other hand, at pH 8, a TAN concentration lower than 50 mg/L makes the MSC curve for NOB shift down (decrease of the FA inhibition) more than the curve for AOB. This situation gives a narrow operational DO range that is impractical, as has been observed in several systems (4, 5), that is, when TAN is lowered too much, nitrate is produced, even though DO is maintained at the same concentration that allowed shortcut nitritation with a higher TAN concentration. Special Applications of MSC Curves. Figure 5 is a contour plot of the DOmin values for TNN and TAN concentrations for pH 8, which is the generally optimal pH for shortcut nitritation. If the actual DO concentration is less than the DOmin value indicated for AOB or NOB, it means that the biomass type should wash out. In most cases, DOmin for the AOB is much smaller than DOmin for NOB. For example, when TAN is 100 mg/L, the system can support the survival of AOB between 10 and 1000 mg/L TNN as long as the DO is greater than 0.18 mg/L. However, the DOmin for NOB is greater than 6 mg/L. Consequently, shortcut nitritation is achievable when the actual DO is between around 0.2 and 6 mg/L, a common operational range. The contours with other pH values can be generated with the equations of the 340
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 1, 2010
FIGURE 6. DOmin curves by pH for shortcut nitritation to provide an input to the ANAMMOX process. We assume that TNN ) TAN for the input to the ANAMMOX treatment. DOmin for NOB is WR at pH 9 because of FA inhibition. The temperature is 30 °C (a, pH 7; b, pH 8; c, pH9). MSC curves; several other examples with pH 7 and 9 are given in the Supporting Information. Recently, a BNR trend is combining shortcut nitritation with anaerobic ammonium oxidation (ANAMMOX), since the combination offers a benefit of avoiding any external carbon source in denitrification (12, 27-29). Figure 6 shows the DOmin values by pH for the special case of linking shortcut nitritation with a typical ANAMMOX treatment. The critical requirement for combining the processes is that the ANAMMOX process needs an influent that has almost same concentrations of ammonium- and nitrite-nitrogen (27). On the basis of typical stoichiometry with biomass synthesis, the NH4+/NO2- molar ratio should be 1:1.3, while it is 1:1
TABLE 4. Comparison with the Minimum DO Concentration in Nitrite Accumulation Systems reference
TAN (mgN/L)
TNN (mgN/L)
Chung et al. (20) Ciudad et al. (4) Blackburne et al. (5) Gali et al. (11) Yamamoto et al. (12) Fux et al. (9)
250 25-100 5-95 350 430 400
170 300-400 5-95 350 570 400
pH
temp (°C)
8 7.8 7.8 6.7a 6.5-7a 6.2-7a
30 25 21 35 25 30
DOminb
NO2/(NOX)
DO (mg O2/L)
duration (days)
AOB
NOB
93% 75% 15%-95% 99% 93% 94%
below 2 1.4 0.4 above 3 5 2-4
800 180 350 160 110 90
0.52 0.15-0.18 0.13-0.15 1.0 0.57-WR 0.45-WR
WR 0.3-WR 0.15-WR WR WR WR
a Uncontrolled. b Washout range (WR) value is shown when TAN or TNN concentration is below SD,min or over SD,max, and thus DOmin numerically approaches infinity.
when biomass synthesis is not considered. The shortcut nitritation process should be controlled to produce a NH4+/ NO2- ratio within this range. For our analysis, we assume that the ratio is 1:1, in which case the abscissa in Figure 6 represents the same concentrations of NH4+-N and NO2--N. The area A indicates when AOB and NOB washout, B indicates only AOB survival, and C represents when both survive. The shortcut system could be maintained in area B. Figure 6 shows that pH 8 is the best for producing the desired shortcut nitritation result for ANAMMOX because it has a large B range and a small C range; with TAN and TNN concentrations around 100-300 mg/L, any DO concentration above 1 mg/L is acceptable at this pH, which makes the reactor operation easy. pH 7 is less desirable, since its C range is relatively large. pH 9 is even less desirable because it has a very small B range and a large A range, which means total system failure. Analysis of Literature Results with MSC Model. We analyzed literature results of several shortcut nitritation processes according to the MSC concept. Table 4 summarizes the results. Values in the left columns are taken directly from the experimental studies, while the right two columns are calculated with the experimental values by eqs 6b and 6c The top three rows show controlled alkaline shortcut nitritation systems, while the following three rows show uncontrolled acidic shortcut nitritation systems. Although not all experimental studies have information on the composition of nitrifying bacteria species, WR values of the estimated DOmin show the possible suppression of AOB and NOB in shortcut nitritation system. For the alkaline systems, the first row (20) shows a WR value for NOB because of the high TAN concentration. Chung et al. (20) showed that almost all of the oxidized N was NO2- and confirmed the suppression of NOB by an antibody test. In the second and third rows, the DOmin values of NOB are WR for high TAN, but can support NOB when TAN is below ∼100 mg/L. This matches with the operating results, in which NO2- did not dominate the oxidized N. Therefore, key to good operation for alkaline system is to maintain a sufficient TAN (or FA) concentration to suppress NOB in the system. For the acidic systems, the WR values of DOmin for NOB support the nitrite-accumulation seen in fourth, fifth, and sixth rows. The DOmin value for AOB are significantly higher in the acidic systems, compared to the alkaline systems. This probably explains why the acidic shortcut systems were maintained at higher DO concentrations. In addition, the acidic systems had higher TNN and TAN concentrations. Park et al. (30) exploited the MSC concept for identifying optimal start-up conditions and achieving stable and low effluent TAN concentrations in a suspended-growth SBNR reactor. The experimental results showed that, whenever the actual TAN concentration was larger than the calculated SD,max for the NOB, the NOB washed out so that shortcut nitrification/denitrification prevailed. Furthermore, the
onset of SBNR was rapid (
