Application of combined process of partial nitritation-anammox using a rotating biological contactor ( PARBC ) to treat ammonium-rich wastewater

Combining the partial Nitritation and Anammox using a rotating biological contactor (PARBC) to remove the ammonium in wastewater was evaluated in this study. The accumulation of Anammox bacteria on the carrier easily obtained after 5 days operating of sequence batch with synthetic wastewater. Then AOB biomass cultivated in PARBC to complete the process of combining two bacteria in the same reactor for completely autotrophic nitrogen removal. After 60 batches of the operation, highest nitrogen removal rate reached 0.33 kg N/m 3 .d with nitrogen removal efficiency is 90% at a concentration of ammonium input of 250 mg N/L. The specific Anammox activity (SAA) of biofilm and suspended sludge in the tank is determined to be 0.298 gN-N2/gVSS/day and 0.0041 gNN2/gVSS/day, respectively. Moreover, the suspended sludge concentration is 17.765 mg MLSS/L. This result showed that Anammox bacteria adapt and grow on the rotating biological carrier; otherwise Anammox bacteria hardly develop in the form of suspended sludge in the tank. This study shows that the PARBCR has great potential to effectively removing ammonium from wastewater with the short startup time.


INTRODUCTION
The CANON (Complete Autotrophic Nitrogen Removal Over Nitrite) process is the combination of partial nitritation and anammox in one reactor [1].This process can be used to remove a high load of ammonia without using external organic carbon [2].According to Strous et al., 1997 [1], Partial Nitritation process in CANON using two autotrophic group of bacteria (aerobic and anaerobic) provided with limited oxygen.The aerobic process occurs by Nitrosomonas and the anaerobic process is by Planctomycete bacteria.Those bacteria consumes ammonia and nitrite to produce nitrogen gas and a small amount of nitrate.The CANON reactor is mixed by air flow.The study points out that CANON granular sludge is formed by Amonia Oxydizing bacteria (AOB) in the surface and Anammox bacteria in the core.
Similar to CANON, the SNAP (Single Stage Nitrogen Removal Using Anammox And Partial Nitritation) was used for nitrogen removal based on partial nitritation and anammox in one reactor.The differences between those two were the using of acrylic medium for attached AOB and anammox [3].Those bacteria are in charge of the transformation of ammonia to nitrogen gas.SNAP has advantages in wastewater treatment practice because the process is more stable and reduce sludge loss.
Both processes have abilities to remove ammonia via two bacteria group AOB and Anammox.The processes can be summarize as the following equation [4]: - (1) This study treated ammonia by the combining process of partial nitrification -Anammox in the same reactor.There were the advantages of both CANON and SNAP processes.Using rotation biological contactor (PARBC) to enrich the biomass as SNAP process while supplying gas to mix suspended sludge, increasing exposure and ensuring the necessary concentration of DO in the reactor as CANON process.By which researching the biomass enriching and Anammox-AOB adaptation to evaluate nitrogen removal effectivity of PARBC model and identify SAA.

PARBC reactor
Firgure 1a presents the schematics of PARBC reactor.The reactor is an acrylic column with (DxH) 300x640 mm, working height of 530 mm, working volume of 35L.The reactor was equipped with a mechanical stirrer to ensure complete mixing.The biomass carrier used in this study is shown in Fig. 1b.
In the start-up phase, DO of the feed wastewater was controlled under 0.5 mg/L using Na 2 SO 3 .pH was maintaned 6.8 -7.0 using HCl and NaHCO 3 [5] .In the main operating phase, DO was controlled 0.8 -1.2 using DO controller (WTW, Germany) and pH was maintained between 7.0 -7.5 by pH controller (WTW, Germany).The PARBC was operated in batch mode.The cycle includes: 15 minutes feed, 45 minutes settling, 15 minutes discharge.The aeration time varies between each tests.

Sludge and wastewater
In the attached phase using synthetic wastewater, 90g granular anammox sludge taken from the IC reactor in Laboratory of FENR -HCMUT was injected into PARBC reactor [5].This granular sludge has VSS/SS = 0.6 and SAA = 0.58 g N 2 /VSS.h.The reactor media was rotated at 10 rpm for mixing and attaching sludge without aeration.The HRT, not including feed, settled and discharge time, was 360 minutes.One cycle is 435 minutes long.At the end of the cycle, the wastewater was sampled and analyzed to measure NH 4 -N, NO 2 -N, NO3--N.Based on the collected data, nitrogen removal efficiency and anammox activity were accessed.When the nitrogen removal efficiency reach 90%, the sludge was sampled and analyzed to measure MLSS, MLVSS, SVI30, SAA.After that, the attached phase was ended.In the main operation phase, the AOB sludge was injected into PARBC reactor.90g of AOB sludge was taken from the pilot PNSBR reactor in Laboratory of FENR -HCMUT.This granular sludge has VSS/SS = 0.76, SAA = 8.88 g N 2 /VSS.h,SVI -40 ml/g.NH 4 Cl (250 mg N/L) was used as substrate for the synthetic water in this phase.The micronutrients comprises of: 1000 mg/L KHCO 3 , 54 mg/L KH 2 PO 4 , 360 mg/L CaCl 2 .2H 2 O, 120 mg/L MgSO 4 .7H 2 O [6].

Attached sludge
The specific activity of Anammox was measured using the pressure method according to Dapena -Mora A., 2006 [7].The Automatic-High-Sensitivity-Gas-Metering-Systems (AHSGMS) consists of an erlen, a pressure meter connected with PC via DAQMaster sofware for continuously monitoring (Firgure 2).The sludge was taken from 16 cm 2 of PARBC media and then washed by phosphate solution (0,14 g/L KH 2 PO 4 ; 0,75 g/L K2HPO4) [7] before feeding into the erlen with 63ml synthetic water.The experiments were conducted with 3 different sludge samples in 3 different compartments of the reactor.The tests were done in room temperature and 150 rpm mixing speed by magnetic stirrer.The SAA values (gN-N 2 /gVSS/day) were calculated based on the nitrogen gas production rate which was determined through the increase of gas pressure in the erlen.

Suspended sludge:
100 ml suspended sludge in PARBC was taken and washed with tap water.0.126 g VSS (dry weight) was feed into the erlen.The experiment were conducted similarly as to the one with attached sludge.

Analysis method
NO 2 -N and NO 3 -N, NH 4 -N, SS, MLSS, MLVSS were determined according to Standard Methods for examination of Water and Wastewater (APHA, 1995).pH and DO were monitored by pH meter (WTW, Germany) and DO meter (WTW, Germany).

Start-up and enrichment of Anammox sludge
The start-up time is 28 days long (28 batches).After the first 5 days, most of the sludge was observed to be attached to the media (Figure 3).The same result could be achieved in SNAP reactor by Dien et al.2013 [8] but with longer operation time of 21 days.This happens due to the rotation of 20 rpm in the reactor which allows the anammox sludge to attach easily to the media compared to the SNAP reactor.Figure 4 shows that the concentrations of ammonia and nitrite in the effluent decrease overtime and stay between 12 -20 mg N/L, corresponding to removal efficiency of over 90%, after 17 days.

Ammonia removal efficiency of PARBC
After start-up phase for attached anammox, the AOB sludge was feed into the reactor.PARBC was operated for 60 batches.The first 5 batches shows low removal efficiency.After 2 days running, most of the AOB sludge was attached to the media.DO was controlled between 0.8 -1.2 mg O 2 /L for AOB activity and reducing anammox inhibition.Figure 5 presents the nitrogen concentration course in the combined phase of anammox and AOB.After 60 batches, the combined partial nitritation and anammox process achieved a ammonia removal of 100% and nitrogen removal of 90%.A brown and occasionally red layer were observed on the attached sludge that suggests good condition of both AOB and anammox.pH in the first 25 batches was controlled at 7 and the next 35 batches was at 7.5.The ammonia removal efficiency of PARBC was high and remained unchanged in both pH levels.The results shows that nitrate concentration of the effluent was always under 15 mg N/L.According to the theory, the CANON process should produce an effluent nitrate of 13% total influent ammonia (32.5 mg N/L in this study).This proves that a part of the produced nitrate was converted by the denitrification bacteria.This means there is a community of denitrification bacteria existed in the reactor along with AOB, NOB and Anammox.However, the nitrite concentration of the effluent remained high (30-60mg N/L).While the ammonia was mostly consumed, the nitrite concentration was still high.This leads to insufficient substrate (ammonia) for anammox bacteria.The solution is to lower the DO concentration in PARBC in order to provide suitable condition for the growth of anammox bacteria.and nitrate increased by 10 mgN/L.This proves that between 4h and 8h, NOB was the dominant specie in the reactor, AOB and anammox activities were constrained due to insufficient of ammonia.In the next experiment (Figure 8), DO was controlled at low level (0.4 -0.8 mg/L).The ammonia concentration decrease from 125 mg/L to 30 mg/L after 8h.The results shows that low DO affect the activity of AOB.The concentration of nitrate was 20 mg/L and remained stable; alkalinity consumed was 720 mg CaCO 3 /L.Table 1 presents the comparison of ammonia and total nitrogen removal rate between this study and others using CANON and SNAP.The result shows that the ammonia and total nitrogen removal rate of PARBC is higher than other study.The PARBC showed great potential in treating ammonia-rich wastewater.

Nitrate removal rate
It was found that the denitrification bacteria existed in PARBC reactor.This experiment was conducted to measure the nitrate removal.Figure 9 shows the nitrate removal rate in 8 hours.In the first 2 hours, the removal rate of the attached sludge was 5.78 mg NO 3 -N /L.h and the suspended sludge was 7.5 times lower.From 2nd hour to 4th hour, the nitrate removal rate decreased due to the lower of substrate (COD).This shows that denitrification bacteria mainly existed on the media due to lower DO in media than in suspended matter.The experiments also shows that the nitrate removal rate depends on the concentration of nitrate which is high in the beginning and rapidly reduce toward the end.The denitrification bacteria in the reactor helps improve the treatment efficiency of nitrogen along with COD removal.

Biomass
The MLVSS feed into the reactor in start-up phase was 477 mg/L (MLVSS/MLSS: 0.6), and increased to 1478 mg/L after 30 days.After the feeding of AOB, the biomass in the reactor was 2185 mg/l (MLVSS/MLSS: 0.76).After 3 months of operation, the biomass concentration in PARBC reactor reach 3.163 mg MLVSS/L.A good growth of both AOB and anammox were observed when using PARBC.

Specific anammox activity (SAA)
The SAA of attached sludge is 0.298 gN-N 2 /gVSS/day and suspended sludge is 0.0041 gN-N 2 /gVSS/day.The anammox activity of attached sludge is higher than suspended because of DO limitation.

CONCLUSIONS
This study shows that anammox and AOB were succesfully attached and adapted in PARBC reactor.The high rate of nitrogen removal of 90% could be easily achieved after 30 days operation.After 60 days running with synthetic wastewater contained 250 mg NH 4 -N/L, the removal rate reached 0.37 kg N/m 3 /d corresponding to an ammonia removal efficiency of 100%.
There is evidence that the AOB, NOB, anammox and denitrification bacteria co-existed in the reactor while AOB and anammox were the main communities to remove nitrogen.The attached anammox community in PARBC is the main contributor to the anammox process with its high activity.
In this study, the concentration of ammonia reached 250 mgN /l and nitrogen loading rate of 0.37 kg N /m 3 .day.Therefore, further studies need to increase the influent concentration of ammonia or nitrogen loading rate to evaluate the nitrogen removing ability of the PARBCR model.

Figure 4 .
Figure 4. Influent and Effluent nitrogen compounds concentration in anammox acclimation period

Figure 5 . 12 Figure 6 .
Figure 5. Influent and Effluent nitrogen compounds concentration in anammox and AOB sludge acclimation period

Figure 7 presents
the nitrogen removal process in 8 hours at DO equals 0.8 -1.2 mg/L.Total alkalinity consumed in one batch was 950 mgCaCO 3 /L.In the first 4 hours, ammonia was completely removed which shows good activity of AOB.At this time, nitrite concentration was 60 mg/L and gradually decreased by 10 mg N/L, A) pH 7 B) pH 7.

Table 1 .
Comparison with other studies