Influence of operating conditions for volatile fatty acids enrichment as a first step for polyhydroxyalkanoate production on a municipal waste water treatment plant

Bioresource Technology 148 (2013) 270–276

Contents lists available at ScienceDirect

Bioresource Technology journal homepage: www.elsevier.com/locate/biortech

Influence of operating conditions for volatile fatty acids enrichment as a first step for polyhydroxyalkanoate production on a municipal waste water treatment plant Timo Pittmann ⇑, Heidrun Steinmetz University of Stuttgart, Institute for Sanitary Engineering, Water Quality and Solid Waste Management, Bandtaele 2, D-70569 Stuttgart, Germany

h i g h l i g h t s  Different flows of a waste water treatment plant were tested as raw material.  Primary sludge yielded the highest amount of VFA and stable VFA composition.  Best operating conditions are: 30 °C, pH 7, retention time 4 d and 25% withdrawal.  Semi-continuous operation method yielded better results than batch operation.  Stable VFA composition could be reached.

a r t i c l e

i n f o

Article history: Received 28 June 2013 Received in revised form 23 August 2013 Accepted 25 August 2013 Available online 4 September 2013 Keywords: Acidification Biopolymer Polyhydroxyalkanoates Primary sludge Volatile fatty acids

a b s t r a c t This work describes the generation of volatile fatty acids (VFAs) as the first step of the polyhydroxyalkanoate (PHA) production cycle. Therefore four different substrates from a municipal waste water treatment plant (WWTP) were investigated regarding high VFA production and stable VFA composition. Due to its highest VFA yield primary sludge was used as substrate to test a series of operating conditions (temperature, pH, retention time (RT) and withdrawal (WD)) in order to find suitable conditions for a stable VFA production. The results demonstrated that although the substrate primary sludge differs in its consistence a stable composition of VFA could be achieved. Experiments with a semi-continuous reactor operation showed that a short RT of 4 d and a small WD of 25% at pH = 6 and around 30 °C is preferable for high VFA mass flow (MF = 1913 mgVFA/(L d)) and a stable VFA composition. Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction Common plastic is derived from petrochemicals based on the limited natural resource petroleum. This is causing serious environmental problems. Beside the exploitation of natural resources the use of plastic is responsible for major waste problems as common plastic is non- or poor biodegradable (UNEP, 2009). Biopolymers present a possible alternative to common plastics. As they are fully biodegradable (Jendrossek and Handrick, 2002; Choi et al., 2004) their use not only allows the preservation of limited resources, but also suits the idea of sustainability. Beside other polymers polyhydroxyalkanoates (PHA), which are biodegradable polyesters accumulated by bacteria under nutrient limited conditions (Nikodinovic-Runic et al., 2013), are a source for bioplastic production. More than 150 component parts of PHA ⇑ Corresponding author. Tel.: +49 71168565852. E-mail address: [email protected] (T. Pittmann). URL: http://www.iswa.uni-stuttgart.de/index.en.html (T. Pittmann). 0960-8524/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.biortech.2013.08.148

have been identified so far (Cavalheiro et al., 2009). The possibility for chemical modification of PHA provide a wide range of material properties and an even wider range of use (Zinn and Witholt, 2005; Akaraonye et al., 2010). However, the main raw material for the biopolymer production are starchy plants like maize (Steinbuechel, 2005), constituting the disadvantages of high land consumption, diminishing food resources as well as problems like leaching of nutrients, input of pesticide and soil erosion (Faulstich and Greiff, 2007). Additionally, bioplastic production is rather expensive, with up to 38% of the costs accounting for the raw material (Lee and Choi, 1997; Choi and Lee, 1999). So far, municipal waste water treatment plants (WWTP) as alternative raw material and biomass source for the PHA production have not been widely investigated, although they offer the opportunity to compensate the disadvantages of the common PHA production using starchy plants. The biological process of PHA production takes place in two steps, which composes the production of volatile fatty acids (VFA) in an anaerobic process and finally the PHA production in

271

T. Pittmann, H. Steinmetz / Bioresource Technology 148 (2013) 270–276

an aerobic process (see also Fig. 1). In contrast to Bengtsson et al. (2008b) and Morgan-Sagastume et al. (2013) the PHA production process described in this work is designed as a side stream process of a municipal WWTP and does not include the treatment of waste water. Therefore the whole process must consider the polymer production only. Chakravarty et al. (2010) showed the possibility to use icecream waste water as alternative source material for the VFA production while Chua et al. (2003) investigatde the effect of pH, sludge-retention time (RT) and acetate concentration on the PHA production from municipal waste water. Diverse authors (Chua and Yu, 1999; Dionisi et al., 2004; Lemos et al., 2006; Reddy et al., 2008) stated that there is a general possibility to produce PHA from activated sludge and Beun et al. (2000) showed that industrial waste water provides a low cost alternative substrate for VFA production. In many of the research projects on PHA production, synthetic waste water was used to gain knowledge about one part of the PHA production or the production’s operating conditions (Albuquerque et al., 2007; Albuquerque et al., 2010; Albuquerque et al., 2011; Bengtsson, 2009; Bengtsson et al., 2010; Choi and Lee, 1997; Dionisi et al., 2005). In contrast, the objective of this research project is to find the most suitable raw material and all operating conditions for the VFA production process using only material flows of a WWTP. This work focuses on the first step of PHA production, the generation of VFAs. At first different raw materials of a municipal WWTP to produce VFAs were observed. Afterwards the influence of operating conditions (temperature, pH, retention time (RT) and withdrawal (WD)) and reactor operation method on VFA production were investigated. Another concern was, how the tested operating conditions or the diversity of the used material flows of a WWTP influence the VFA composition and consequently the kind of PHA produced. As there is a variation in the composition of the used material flows (sludges) of a WWTP, it is of particular importance to observe their influence on VFA production and composition. 2. Methods 2.1. Experimental set-up The overall production process to produce biopolymers from municipal waste water is displayed in Fig. 1. This work covers ‘‘(1): Acidogenic fermentation’’ only. Therefore anaerobic reactors of different sizes (4 L, 15 L) were operated as batch reactors or as semi-continuous reactors. While the batch operation is defined as a one-time substrate filling at the beginning of the experiment

with no withdrawal and refill during the test, a semi-continuously operation method allows to introduce and withdraw substrate to the reactor. Semi-continuously operation means that the amount of substrate, which has to be changed is withdrawn and refilled at once. There was no sedimentation or biomass recirculation in all tests. Therefore the hydraulic retention time equals the sludge age and both will be referred hereinafter as RT. As most of the results influence the following tests, a chronological test order was implemented as follows: 1. Selection of raw material. 2. Evaluation of operating conditions including. (a) Selection of suitable pH-level. (b) Evaluation of a RT range. (c) Selection of a suitable combination of RT and withdrawal (WD). The raw material selection has the highest priority. Therefore it was investigated at first. The optimisation of the operating conditions for the VFA production was examined afterwards with the most appropriate raw material found.

2.1.1. Selection of raw material For raw material selection continuously stirred batch reactors with a volume of 4 L were used. Four different sludges, namely primary sludge (average total solid (aTS) = 43 g/L), excess sludge (aTS = 10 g/L), a one to one mixture of primary- and digested sludge (aTS = 37.5 g/L) and a one to one mixture of excess- and digested sludge (aTS = 21 g/L) from a municipal WWTP were analysed. Thereby digested sludge was only used as inoculum for the anaerobic process in order to find out, if it could accelerate the process. All named sludges were investigated under four different conditions: pH controlled at pH = 6, without pH-control and each at around 20 °C or around 30 °C reactor temperature. In summary 16 different tests were performed. The reactors were filled at the beginning of the experiments and samples of 50 mL were retrieved every day to determine the VFA concentration and composition. To achieve the selected temperature the reactors were situated in temperature-controlled rooms. For pH-controlled tests, the pH-value was measured by a mobile pH meter (WTW pH340i) and adjusted with NaOH by hand twice a day. The test duration for all experiments was 18–20 d. A sample of all tested sludges was taken before and after the tests to determine COD, TKN and total P.

NaOH

mineral nutrients PHA accumulating bacteria

centrifuge primary sludge

VFA containing water anaerobic reactor

aerobic reactor

sampling or excess sludge

(2a): Biomass accumulation

biomass

biomass with high PHA storage capacity

(1): Acidogenic fermentation aerobic reactor sampling dewatering and chemical extraction of PHA

biomass with high PHA concentration

(2b): PHA production

Fig. 1. Experimental set-up for the biopolymer production from waste water.

272

T. Pittmann, H. Steinmetz / Bioresource Technology 148 (2013) 270–276

2.1.2. Evaluation of operating conditions As the influence of fermentation temperature was already observed sufficiently during the selection of the raw material, three additional operating conditions (pH, RT, WD) were investigated in continuously stirred reactors with a volume of 15 L. For all experiments primary sludge from a municipal WWTP was used as raw material and a sample was retrieved prior to the experiments to determine COD, TKN and total P. All reactors were placed in a temperature-controlled room at around 30 °C. pH was monitored at all times via a Metrom Profitrode pH probe and automatically adjusted with NaOH during all tests. Some experiments to identify the best operating conditions were performed as batch-tests (pH, pre-RT), while the reactors were operated semicontinuously to discover the best RT and WD. For the batch tests all reactors were filled with primary sludge at the beginning and samples of 100 mL were retrieved every day to determine the VFA concentration and composition. The batch test period was 18–20 d long. Bengtsson et al. (2008b) stated that a pH level below 5 inhibits the VFA production. Chen et al. (2013) yielded the highest VFA production at pH = 9, while Mengmeng et al. (2009) reached the best VFA results at pH = 11. In consequence a range of pH-levels (pH = 6, 6.5, 7, 8, 10) was tested. Bengtsson et al. (2008b) described the influence of RT on the VFA production. To find an approximate RT for semi-continuously operation, so called pre-RT experiments were conducted: the reactor was filled in three subsequent steps so that already adapted microorganisms were able to ferment fresh substrate. In the beginning the reactor was filled with 3 L primary sludge. After a starting phase of 10 d a second load of 3 L primary sludge was introduced. Samples of 100 mL were retrieved every day to determine the VFA concentration and composition. When the VFA concentration reached its maximum, the final 3 L primary sludge were introduced to the reactor. Again VFA concentration and composition was observed via samples of 100 mL until a maximum was reached. For the semi-continuously operated tests all reactors were filled with primary sludge at the beginning and operated under pH-controlled conditions at pH = 6. After a starting phase of 10 , to accumulate VFAs, the semi-continuous operation phase began, for which a certain amount of the sludge in the reactor was exchanged. Samples of 100 mL were retrieved to analyse the VFA concentration and composition for about 40 d with a RT of 4 d, 6 d and 8 d, each with 25% and 50% WD. Additionally a 75% WD was performed with a RT of 4 d. A RT of 2 d was also tested with a WD of 50%. RT and WD are related factors, e.g. a RT of 4 d was used when 25% of the sludge was exchanged every day, 50% every second day or 75% every third day.

2.3. Calculation of operating conditions For a better comparability all results regarding VFA concentrations, COD, TKN and total P were converted into mg/L. In this study the concentration of VFA in terms of mg/L is:

VFA ¼

Ac þ Pro þ Bu

ð1Þ

The degree of acidification (DA) was calculated according to Bengtsson et al. (2008a) as shown in Eq. 2. As VFA results are given in mgVFA/L they have to be converted into COD units as shown in Eq. 3.

DA ¼

VFA CODS

in

mgCOD =L mgCOD =L

ð2Þ

with CODS = COD of Substrate at the start.

CODVFAi ¼

conc:VFAi  oxygen demandi molar mass VFAi

ð3Þ

with i = Ac, Pro, Bu. For a better comparability regarding the different RT and WD, the average of the VFA concentration during the test period was calculated and converted into a mass flow. The conversion also eliminates the reactors size and can hence be considered as average VFA mass flow per day and litre, which will be referred to mass flow (MF) hereinafter (Eq. 4).

MFVFA ¼

av: VFA conc: RT

mgVFA Ld

in

ð4Þ

As result reactor sizes and retention times can be compared. 3. Results and discussion 3.1. Raw material Table 1 displays the results of the performed investigations ordered by degree of acidification. Primary sludge performed best under three out of four conditions and yielded by far the best DA with 31% at 30 °C under pH-controlled conditions. The second best carbon source, a one to one mixture of primary and digested sludge at 20 °C under pH-uncontrolled conditions, achieved only a DA of

Table 1 Degree of acidification and VFA composition in dependence of substrate and operating conditions (batch-tests, 4 L).

2.2. Analytical procedures COD, TKN, total P and total solids (TS) were determined according to standard methods. The concentration and composition of volatile fatty acids, namely formate (Fo), acetate (Ac), propionate (Pro) and butyrate (Bu), were detected by high performance liquid chromatography (HPLC). Therefore the sample was acidulated to pH = 2 and filtered through 0.45 lm membrane filter. Afterwards HPLC detection was performed using a HP1100 chromatographer equipped with an UV detector and a Varian Metacarb 87H column. Sulphuric acid 0.05 M was used as eluent at a flow rate of 0.6 mL/min. The detection wavelength was 210 nm. Volatile fatty acid’s concentration was calibrated using 4–4000 nmol standards. As the results of formic acid detection was below detection point for all except one test, formic acid is not shown in the VFA composition.

X

a

Carbon source

pH

Temp. (°C)

(d)

(%)

Ac/Pro/ Bu (%)

Primary sludge Primary sludge Primary sludge Primary-/digested sludge Primary sludge Primary-/digested sludge Excess sludge Excess sludge Primary-/digested sludge Excess sludge Excess sludge Primary-/digested sludge Excess-/digested sludge Excess-/digested sludge Excess-/digested sludge Excess-/digested sludge

6a 6a 4,6 7

30 20 30 20

9 7 10 14

31 14 14 14

52/48/0 56/44/0 41/59/0 79/21/0

4,5 7,5

20 30

15 14

13 12

42/58/0 84/16/0

7 6,5 6a

30 20 30

5 4 5

10 8 7

59/20/21 60/20/20 75/25/0

6a 6a 6a

20 30 20

7 5 2

6 6 3

24/76/7 67/33/0 57/43/0

6a 8 7,5 6a

30 30 20 20

4 3 7 2

3 3 2 1

100/0/0 76/0/24 100/0/0 100/0/0

Marks conditions pH-controlled.

Max. conc.

DA

273

T. Pittmann, H. Steinmetz / Bioresource Technology 148 (2013) 270–276

14%. In five out of eight experiments pH-uncontrolled conditions reached a higher DA. Therefore a fermentation without pH control should be considered for all fermentation raw materials. However, primary sludge yielded better DAs with pH control at both investigated temperatures. At 30 °C the DA of primary sludge was twice as high than without pH control. To determine the effect of using digested sludge as inoculum on VFA production the day at which the maximum VFA concentration (Max. conc.) was reached is shown in Table 1. In six out of eight combinations an inoculum with digested sludge decreased the time at which the maximum VFA concentration was determined in the reactor. Primary sludge at 30 °C under pH-controlled conditions attained its maximum after 9 d without and after 5 d with added digested sludge. Although the maximum concentration was reached faster with the inoculum, there was a significant decline in DA in seven out of eight tested combinations, which can be explained by the lower substrate availability of digested sludge. With Primary sludge at 30 °C under pH-controlled conditions the DA declined from 31% to only 7%. Therefore the digested sludge should not be considered as an inoculum for the VFA production. Table 1 also shows the composition of the VFA. The results varied strongly between 24/76/7 (%Ac/%Pro/%Bu) and 100/0/0 depending on the used raw material. Primary sludge produced none butyric acid and acetic and propionic acid in nearly two equal sections. Excess sludge on the other hand produced up to 21% butyric acid, while the one to one mixture of primary and digested sludge produced the most acetic acid (up to 84%) of all tested raw materials. The results show that the raw material has a major influence on the VFA composition. As the use of primary sludge resulted in highest DA and showed only small variations in VFA composition under the tested conditions, it was chosen as raw material and used in all further tests. Beside the ability to produce VFAs, primary sludge has other advantages as raw material for the PHA production. As primary sludge is a mixture of organic material, water and fermenting microorganisms no longsome biological adaptation-phase or biomass recirculation for the fermentation process was needed. During all experiments primary sludge showed nutrient limited conditions, with a C:N:P-ratio in a range of 100:2:0.5 to 100:3:0.8. There was also a measurement of the C:N:P-ratio in the VFA containing water at the end of the fermentation process. In comparison to the primary sludge only small changes in the C:N:P-ratio were detected. The VFA containing water also showed nutrient limited conditions with a C:N:P-ratio in a range of 100:2.6:0.5 to 100:3.6:0.7. This is of particular significance given that nutrient limited conditions are essential for the later PHA production (Serafim et al., 2004; Albuquerque et al., 2007).

VFA-concentration in mg/L

20000

3.2. Temperature Aim of the investigations at two temperature levels was to ascertain that the VFA production at ambient temperature (20 °C) can reach the same VFA production compared with heating the sludge. In six out of eight tested combinations a temperature increase from 20 °C to 30 °C caused a higher VFA production as shown in Table 1. The experiment confirmed the results of Mengmeng et al. (2009), who stated that the VFA concentration increases with higher fermentation temperature. Using primary sludge as raw material (under pH-controlled conditions) the temperature change from 20 °C to 30 °C caused a DA increase from 14% to 31%. The general assumption that the acidification rate is higher at 30 °C than at 20 °C could not be confirmed. Only three out of eight tested combinations reached their VFA maximum at 30 °C in a shorter span of time than at 20 °C. Four out of them even reached their VFA maximum at 20 °C in a shorter span of time than at 30 °C. The results can be seen in Table 1. Primary sludge under pH-controlled conditions obtained its VFA maximum after 7 d at 20 °C and after 9 d at 30 °C. Nevertheless, the fact that the DA of primary sludge under pH-controlled conditions at 30 °C was twice as high as the DA at 20 °C is all the more important as the VFA production at 30 °C lasted only about 30% longer. The variation of temperature has a wide range of influence on the VFA composition, depending on the used substrate. While the mixture of primary-/digested sludge produced just more acetic acid at 30 °C (%Ac/%Pro/%Bu = 75/25/0) than at 20 °C (57/43/0), a complete reversal of the percentage due to the temperature change using excess sludge from 24/67/7 at 20 °C to 67/33/0 at 30 °C was observed. As primary sludge was already chosen as substrate for the optimisation tests, only its VFA composition was of interest for further tests. However, in the case of primary sludge the temperature change investigated resulted only in marginal changes in the VFA composition. In consequence a temperature around 30 °C for the further experiments was considered as reasonable. 3.3. pH As illustrated in Fig. 2 no big difference in the maximum VFA concentration between pH = 6 and pH = 8 was observed. A pH value of 7 yielded the highest result with 18,286 mgVFA/L after a RT of 10 d. The fermentation at pH = 10 reached significantly worse results with a maximum of 10,050 mgVFA/L only at 18 d retention time. This is in contrast to the results of Chen et al. (2013) showing the best result at pH = 9 with excess sludge and food waste as source material and Mengmeng et al. (2009) yielding the highest result at pH = 11 with excess sludge as raw material.

pH = 6 pH = 6,5 pH = 7 pH = 8 pH = 10

18000 16000 14000 12000 10000 8000 6000 4000 2000 0

0

2

4

6

8

10 time in d

12

14

16

Fig. 2. VFA concentration in dependence of pH (batch-tests, 15 L).

18

20

274

T. Pittmann, H. Steinmetz / Bioresource Technology 148 (2013) 270–276

Table 2 VFA composition and DA in dependence of pH or RT and WD. Batch

a

Semi-continuous

pH

Ac/Pro/Bu (%)

DA (%)

pH

RT and WD (d and %)

Ac/Pro/Bu (%)

DA (%)

6 6.5 7 8 10a

45/51/4 37/61/2 28/72/0 60/37/3 45/31/8

29 29 39 29 14

6 6 6 6 6 6 6

4 4 4 6 6 8 8

49/38/13 46/47/7 48/45/7 51/38/11 46/48/6 46/41/13 43/49/8

15 14 14 22 17 20 19

and and and and and and and

25 50 75 25 50 25 50

Missing to 100% is formic acid.

Although pH = 7 yielded the best result, methane production turned out to be an issue at this pH-value. After about 15 d the acetate concentration was falling rapidly and the overall VFA concentration was less than 44% of the maximum after 18 d (Fig. 2). During this period more than 15 vol.% methane was detected in the reactor via gas measurement. To prevent methanogenic conditions a pH-level of 6 is to be kept (Bischofsberger et al., 2005). Consequently further investigations were performed under pH = 6 although it produced about 12% less VFA. The variation of pH-level showed a strong influence on the VFA composition. Changing the pH from 6 to 7 within the batch experiments caused a constant decrease in the acetic acid ratio as shown in Table 2. In the same tests, the propionic acid ratio increased, while the butyric acid ratio decreased to zero. At pH = 8 conditions a reverse trend was observed. With 60% the maximum acetic acid ratio as well as the minimum propionic acid ratio (37%) was detected, while butyric acid was produced in small amounts (3%) again. In comparison to pH = 8, a reduction of acetic acid and propionic acid production was detected at the highest tested pH-level (pH = 10), while the butyric acid ratio increased to the highest level (8%) observed. In contrast to the other pH-values tested, formic acid was produced at pH = 10 with a ratio of 16%. Albuquerque et al. (2007) showed that the PHA production and composition is directly influenced by the VFA composition. Therefore a change in pH-level is not only effecting the VFA production, but also important regarding the material properties of the PHA. 3.4. pre-RT Studies of other authors showed a wide variation of optimal RTs between a few hours (Albuquerque et al., 2010; Bengtsson et al., 2008b) and several days (Rhu et al., 2003; Bengtsson et al.,

transitional phase

2008a; Mengmeng et al., 2009) depending on the used substrate. As primary sludge has not been used as substrate yet, pre-RT tests were necessary to find an approximate RT. The results of the test showed that a shorter RT was needed in semi-continuously operated reactors than in batch reactors, as the maximum VFA production was reached in a shorter time, if already adapted microorganisms were in the reactor. While the maximum VFA concentration was reached after about 10–18 d in the previous batch-test (see Fig. 2), it took 5 d only after the second load of primary sludge was introduced to the reactor. After the third load of subtrate was filled into the reactor it took only 3 d to achieve the maximum VFA concentration. As there was no continuous addition of substrate and withdrawal of mixed liquid in the reactor during this test, the results gave only a rough estimate of the best RT. In consequence RTs between 2 d and 8 d were tested within further experiments. 3.5. RT and WD To get an idea about how much adapted bacteria are needed in the reactor to produce the most VFAs a wide range of RT and WD was tested. RT = 2 d and WD = 50% yielded poor results and after 10 d of semi-continuous operation the test was shut down. Therefore these results are not shown. Further results ranged in a broad band between 5000 mgVFA/L and 10,000 mgVFA/L. Obviously the VFA production with short RTs and small WDs fluctuated less than using long RTs and large WDs, what can be explained by the changing composition of the introduced primary sludge. After a short rainfall the primary sludge contains high amounts of organic material, because of a flush of the sewer system, while during long rain periods it is highly diluted with low organic material. Smaller WD stabilise the fermentation process in such cases, for only little material is turned over and the reactor is less sensitive to heterogeneous primary sludge input. Fig. 4 shows the average VFA concentration over a total time of 40 d for all used combinations. Both, RT and WD influenced the VFA production. With higher WD the VFA concentration was decreasing at all tested RTs. The highest overall VFA concentration was reached at a RT of 6 d with a WD of 25%. Longer and shorter RTs (with a WD of 25% also) yielded lower VFA concentrations. For a later PHA production in a second stage a high VFA mass flow is preferable to a high VFA concentration, which could lead to a substrate inhibition (Albuquerque et al., 2010). Therefore the MF was calculated on Eq. 4 and illustrated in Fig. 4. A RT = 4 d and WD = 25% yielded the top mass flow with MF = 1913 mgVFA/ (L d) at a VFA concentration of 7653 mgVFA/L on average.

semi-continuous phase

Butyrate

Propionate Acetate

Fig. 3. Development of the VFA concentration at RT = 4 d, WD = 50% (semi-continuous-tests, 15 L)

T. Pittmann, H. Steinmetz / Bioresource Technology 148 (2013) 270–276

275

Fig. 4. Average VFA concentration and VFA mass flow in dependence of RT and WD (semi-continuous-tests, 15 L).

The variation of RT influenced the VFA composition only marginal as shown in Table 2. Acetic acid and propionic acid were produced in nearly the same amount, while butyric acid was only a small part. Nevertheless, the variation of WD had an effect. With a WD of 25% the fraction of propionic acid was about 20% smaller throughout all tests, compared to a WD of 50% and 75%, while the butyric acid ratio was about twice as much. The acetic acid ratio with a WD of 25% was slightly higher than for any other WDs. As a stable VFA composition is necessary for high quality PHA production the possible fluctuation of the VFA composition (due to changing composition of the introduced primary sludge) during the whole test period is of importance. Fig. 3 is exemplary for all semi-continuously operated tests and shows the VFA concentration at RT = 4 d, WD = 50% divided into the three tested VFAs during the whole test period of 44 d. After a starting phase of 10 d (not shown in the figure) the semi-continuous operation began. Due to the change in the operation method a transition phase with a decrease in VFA concentration was observed for the first six days of semi-continuous operation. Fluctuations in the VFA concentration between day six and day 44 were due to the changing composition of the introduced primary sludge. Although a fluctuation in VFA concentration was observed, there were only minor changes in the VFA composition. Thus it was possible to show that the variability of the raw material primary sludge did not affect the VFA composition. 3.6. Operation method There are two possibilities to produce VFA from primary sludge. Batch reactors or semi-continuously operated reactors. As mentioned in Chap. 2.1.2, the substrate-, temperature-, pH- and pre-RT

tests were conducted as batch tests and yielded by far the highest VFA concentrations with a maximum of over 18,000 mgVFA/L. The peak concentrations were reached after about 10–18 d (Fig. 2). Although these tests do not had the best MF (with a maximum of 1829 mgVFA/(L d) at pH = 7, RT = 10 d, WD = 100%), they have some advantages compared to the semi-continuously operated experiments. After about 15 d the risk of methane production and in comparison a massive VFA loss in the semi-continuously operated reactors increased (see also Chap. 3.3). As the maximum working period of a batch reactor equals the RT and the RT at pH = 7 is about 10 d, the risk of methane production during fermentation of primary sludge in a batch reactor was barely existing. One of the disadvantages of the batch reactors was the very high VFA concentration, which has to be diluted or could lead to a substrate inhibition (Albuquerque et al., 2010), depending on the solid concentration (and thus the sludge load) in the second stage reactor. The second disadvantage is the high RT of 10 d and large WD of 100%. To withdraw VFAs every day, at least 10 batch reactors, inclusively their peripheral equipment are needed. Semi-continuously operated reactors allow a daily VFA supply with one unit only. As mentioned in Chap. 3.5, a large WD is more sensitive to variances in the primary sludge composition and therefore fluctuations in the VFA production would be higher when using batch reactors. The operation method of the reactor influenced the VFA composition as shown in Table 2, due to the associated change in WD. Under pH = 6 conditions the VFA composition in the batch reactor (WD = 100%) consisted of 45/51/4 (%Ac/%Pro/%Bu). The semi-continuous reactor produced 49/38/13 (with RT = 4 d and WD = 25%) and therefore less propionic acid in favour of butyric acid. This is of importance as Marang et al. (2013) showed that the carbon

276

T. Pittmann, H. Steinmetz / Bioresource Technology 148 (2013) 270–276

uptake rate and PHA production using butyrate as source material instead of acetate was higher and concluded that fermentation should be optimized on butyrate production. On the other hand Venkata Mohan and Venkateswar Reddy (2013) stated that the most important individual factor to enhance PHA production is the microenvironment,and then the pH. VFA composition is less important to the PHA production, but still has an influence on the kind of PHA produced. To produce PHA on a continuously basis, a constant supply with VFAs is necessary. Therefore the semi-continuous operation method is preferable, particularly with regards to a constant and stable VFA production, higher butyrate production and investment and operational costs. The operation method strongly influenced the DA, as illustrated in Table 2. During batch operation the DA was much higher, with an exception at pH = 10. The batch test at pH = 6 resulted in a DA of 29%, while the semi-continuous experiments (also pH = 6) yielded their highest DA of 22% at a RT = 6 d and WD = 25%. Contrary RT and WD with the best MF (RT = 4 d and WD = 25%) converted only up to a DA of 15%. For the upcoming use of the fermented primary sludge the DA is of interest, too. On the one hand a high DA is desirable for VFA production, on the other hand it leads to a low COD in the separated biomass, which limits its opportunities for subsequent utilisation. The possibility to reuse the biomass in other parts of the WWTP is going to be determined in future tests. 4. Conclusion From the results, it could be concluded that it is possible to produce high amounts of VFAs with a stable VFA composition on a WWTP. Using different raw materials shows a strong influences on DA and VFA composition. The VFA production and composition is strongly influenced by a pH-level change in the reactor. A semicontinuous operation method of the reactor with a short RT and small WD is preferable. With primary sludge as raw material no biomass recirculation for the fermentation process is needed. Acknowledgements We thank the WILLY-HAGER-STIFTUNG, Stuttgart for funding the research project. Our thanks go out to Florian Leo for his great help in reading, correcting, formulation and spell checking. References Akaraonye, E., Keshavarz, T., Roy, I., 2010. Production of polyhydroxyalkanoates: the future green materials of choice. Journal of Chemical Technology & Biotechnology 85, 732–743. Albuquerque, M.G.E., Concas, S., Bengtsson, S., Reis, M.A.M., 2010. Mixed culture polyhydroxyalkanoates production from sugar cane molasses: the use of a 2stage cstr system for culture selection. Bioresource Technology 101, 7112– 7122. Albuquerque, M.G.E., Eiroa, M., Torres, C., Nunes, B.R., Reis, M.A.M., 2007. Strategies for the development of a side stream process for polyhydroxyalkanoate (pha) production from sugar cane molasses. Journal of Biotechnology 130, 411–421. Albuquerque, M.G.E., Martino, V., Pollet, E., Avrous, L., Reis, M.A.M., 2011. Mixed culture polyhydroxyalkanoate (pha) production from volatile fatty acid (vfa)rich streams: effect of substrate composition and feeding regime on pha productivity, composition and properties. Journal of Biotechnology 151, 66–76. Bengtsson, S., 2009. The utilization of glycogen accumulating organisms for mixed culture production of polyhydroxyalkanoates. Biotechnology and Bioengineering 104, 698–708. Bengtsson, S., Hallquist, J., Werker, A., Welander, T., 2008a. Acidogenic fermentation of industrial wastewaters: effects of chemostat retention time and pH on volatile fatty acids production. Biochemical Engineering Journal 40, 492–499. Bengtsson, S., Pisco, A.R., Reis, M.A.M., Lemos, P.C., 2010. Production of polyhydroxyalkanoates from fermented sugar cane molasses by a mixed culture enriched in glycogen accumulating organisms. Journal of Biotechnology 145, 253–263.

Bengtsson, S., Werker, A., Christensson, M., Welander, T., 2008b. Production of polyhydroxyalkanoates by activated sludge treating a paper mill wastewater. Bioresource Technology 99, 509–516. Beun, J.J., Paletta, F., Van Loosdrecht, M.C.M., Heijnen, J.J., 2000. Stoichiometry and kinetics of poly-b-hydroxybutyrate metabolism in aerobic, slow growing, activated sludge cultures. Biotechnology and Bioengineering, 379–389. Bischofsberger, W., Dichtl, N., Rosenwinkel, K.H., Seyfried, C.F., Bohnke, B., 2005. Anaerobtechnik. Springer Verlag, Berlin Heidelberg. Cavalheiro, J.M.B.T., de Almeida, M.C.M., Grandfils, C., da Fonseca, M.M.R., 2009. Poly(3-hydroxybutyrate) production by cupriavidus necator using waste glycerol. Process Biochemistry 44, 509–515. Chakravarty, P., Mhaisalkar, V., Chakrabarti, T., 2010. Study on polyhydroxyalkanoate (pha) production in pilot scale continuous mode wastewater treatment system. Bioresource Technology 101, 2896–2899. Chen, H., Meng, H., Nie, Z., Zhang, M., 2013. Polyhydroxyalkanoate production from fermented volatile fatty acids: effect of pH and feeding regimes. Bioresource Technology 128, 533–538. Choi, G.G., Kim, H.W., Rhee, Y.H., 2004. Enzymatic and non-enzymatic degradation of poly (3-Hydroxybutyrate-co-3-Hydroxyvalerate) copolyesters produced by alcaligenes sp. MT-16. The Journal of Microbiology 42 (4), 346–352. Choi, J., Lee, S.Y., 1997. Process analysis and economic evaluation for poly(3hydroxybutyrate) production by fermentation. Bioprocess Engineering 17, 335. Choi, J., Lee, S.Y., 1999. Factors affecting the economics of polyhydroxyalkanoate production by bacterial fermentation. Applied Microbiology and Biotechnology 51, 13–21. Chua, A.S.M., Takabatake, H., Satoh, H., Mino, T., 2003. Production of polyhydroxyalkanoates (pha) by activated sludge treating municipal wastewater: effect of pH, sludge retention time (srt), and acetate concentration in influent. Water Research 37, 3602–3611. Chua, H., Yu, P.H.F., 1999. Production of biodegradable plastics from chemical wastewater – a novel method to reduce excess activated sludge generated from industrial wastewater treatment. Water Science and Technology 39, 273–280. Dionisi, D., Carucci, G., Papini, M.P., Riccardi, C., Majone, M., Carrasco, F., 2005. Olive oil mill effluents as a feedstock for production of biodegradable polymers. Water Research 39, 2076–2084. Dionisi, D., Majone, M., Papa, V., Beccari, M., 2004. Biodegradable polymers from organic acids by using activated sludge enriched by aerobic periodic feeding. Biotechnology and Bioengineering 85, 569–579. Faulstich, M., Greiff, K.B., 2007. Klimaschutz durch biomasse, ergebnisse des srusondergutachtens. Jendrossek, D., Handrick, R., 2002. Microbial degradation of polyhydroxyalkanoates. Annual Review of Microbiology 56, 403–432. Lee, E.Y., Choi, C., 1997. Biosynthesis and biotechnological production of degradable polyhydroxyalkanoic acid. Biotechnology and Bioprocess Engineering 2, 1–10. Lemos, P., Serafim, L., Reis, M.A.M., 2006. Synthesis of polyhydroxyalkanoates from different short-chain fatty acids by mixed cultures submitted to aerobic dynamic feeding. Journal of Biotechnology 122, 226–238. Marang, L., Jiang, Y., van Loosdrecht, M.C.M., Kleerebezem, R., 2013. Butyrate as preferred substrate for polyhydroxybutyrate production. Bioresource Technology 142, 232–239. Mengmeng, C., Hong, C., Qingliang, Z., Shirley, S.N., Jie, R., 2009. Optimal production of polyhydroxyalkanoates (pha) in activated sludge fed by volatile fatty acids (vfas) generated from alkaline excess sludge fermentation. Bioresource Technology 100, 1399–1405. Morgan-Sagastume, F., Valentino, F., Hjort, M., Cirne, D., Karabegovic, L., Geradin, F., Dupont, O., Johansson, P., Karlsson, A., Magnusson, P., Alexandersson, T., Bengtsson, S., Majone, M., Werker, A., 2013. Biopolymer production from sludge and municipal wastewater treatment, in: 1st International IWA Conference on Holistic Sludge Management. Nikodinovic-Runic, J., Guzik, M., Kenny, S., Babu, R., Werker, A., O’Connor, K., 2013. Carbon-rich wastes as feedstocks for biodegradable polymer (polyhydroxyalkanoate) production using bacteria. Advances in Applied Microbiology 84, 139–200. Reddy, S.V., Thirumala, M., Reddy, T.V.K., Mahmood, S.K., 2008. Isolation of bacteria producing polyhydroxyalkanoates (pha) from municipal sewage sludge. World Journal of Microbiology and Biotechnology 24, 2949–2955. Rhu, D.H., Lee, W.H., Kim, J.Y., Choi, E., 2003. Polyhydroxyalkanoate (pha) production from waste. Water Science and Technology 48 (2), 221–228. Serafim, L.S., Lemos, P.C., Oliveira, R., Reis, M.A.M., 2004. Optimization of polyhydroxybutyrate production by mixed cultures submitted to aerobic dynamic feeding conditions. Biotechnology and Bioengineering 87 (2), 145– 160. Steinbuechel, A., 2005. Angewandte Mikrobiologie, Biopolymere und Vorstufen. Springer-Verlag, Heidelberg, pp. 373–396. UNEP, 2009. United Nations Environment Programme (UNEP): Marine Litter – A Global Challenge, 2009, Nairobi, Kenya. Venkata Mohan, S., Venkateswar Reddy, M., 2013. Optimization of critical factors to enhance polyhydroxyalkanoates (PHA) synthesis by mixed culture using Taguchi design of experimental methodology. Bioresource Technology 128, 409–416. Zinn, M., Witholt, B., 2005. Tailored material properties of poly-hydroxyalakanotes through biosynthesis and chemical modification. Advanced Engineering Materials 7, 408–411.