Treatment of terephthalic acid plant wastewater with an

anaerobic fixed film reactor

A. Noyola, Hervé Macarie, J.-P Guyot

To cite this version:

A. Noyola, Hervé Macarie, J.-P Guyot. Treatment of terephthalic acid plant wastewater
with an anaerobic fixed film reactor. Environmental Technology, 1990, 11 (3), pp.239 - 248.
�10.1080/09593339009384862�. �hal-01737867v2�

HAL Id: hal-01737867

https://hal.science/hal-01737867v2
Submitted on 7 Apr 2018

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est

archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents
entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non,
lished or not. The documents may come from émanant des établissements d’enseignement et de
teaching and research institutions in France or recherche français ou étrangers, des laboratoires
abroad, or from public or private research centers. publics ou privés.
Authors’ version of the article published in Environmental Technology, 11 (3), 239-248 (1990)
and available a DOI: 10.1080/09593339009384862

TREATMENT OF TEREPHTHALIC ACID PLANT WASTEWATER WITH AN

ANAEROBIC FIXED FILM REACTOR

A. Noyola1, H. Macarie, J.-P. Guyot2

Universidad Autónoma Metropolitana, Departamento de

Biotecnología, A.P. 55-535, Iztapalapa, 09340 México D.F.,
México
1
Instituto de Ingeniería UNAM, A. P. 70472, Coyoacan,
04510 México, D.F., México.
2
Mission ORSTOM, Homero 1804-404, los Morales 11510,
México, D.F., México.

ABSTRACT

Anaerobic treatment of terephthalic acid plant wastewater

with a lab-scale downflow tubular fixed film reactor
achieved 75% of COD removal with a HRT of 3.4 days at
33°C. The raw wastewater from a petrochemical plant was
settled and neutralized prior to reactor feeding. The
responses to shock loads and to periods without feeding
were satisfactory, and no inhibitory effects were noticed.
A primary settling - anaerobic - aerobic process
configuration is proposed as a highly efficient, energy
saving alternative to the conventional aerobic process.

INTRODUCTION

Polymer grade terephthalic acid (1,4-benzenedicarboxylic

acid) is essentially used to make poly(ethy1ene
terephthalate), the basic polymer employed in the
manufacture of polyester textile fibers, polyester films
(X-ray films, microfilms, audio and video recording
tapes), food packaging, beverage bottles, adhesives,
coatings etc. (1). This compound is produced in the USA,
Japan, Western and Eastern Europe, Latin America and
Southeast Asia. In 1986, 4.2 millions tons were
manufactured (2), and since this time, the production
capacity of several plants have been increased and new
plant projects have been announced or realized.

The wastewater from factories producing terephthalic acid

1
is generally treated by three stage activated sludge
process (3). This arrangement allows good total COD
removal (more than 90%) and it has proved to be resistant
to shocks of twice the normal organic load when treating a
synthetic wastewater simulating an effluent from
terephthalic acid plants (3). In addition, it has been
reported that terephthalate and p-toluate (both at COD =
1000 mg.l-1), which are the principal aromatic
contaminants of this effluent, were degraded in fed batch
activated sludge reactors to 96% and 88% respectively (4).

However, the aerobic treatment of this kind of wastewater

presents two principal disadvantages: high O2 (energy)
requirements and poor settling sludge (3) or even bulking
sludge (5). Furthermore, the terephthalic acid plant
effluent has deficiencies in nitrogen and phosphorus;
these elements must be added at a ratio of 200:5:1
(C0D:N:P) for aerobic treatment, which means important
ammonia and phosphoric acid consumption.

Compared with aerobic treatment, the anaerobic process has

lower nutrient requirements and sludge production, and it
may be an energy producer. In addition, if second or third
generation reactors are employed for high strength
wastewater treatment, plant size can be smaller than in
aerobic systems. Anaerobic reactors are mainly applied to
food processing industry effluents which are readily
biodegradable. Nevertheless, various studies prove the
feasibility of anaerobic digestion in treating chemical
wastewaters such as those from coal gasification, paint
stripping operations, refinery sour water stripper bottoms
(6), coal liquefaction (7), coal coking (8), manufacture
of ethylene and propylene from naphta (9) production of
polyolefins, pharmaceutical intermediates and acrylic
resins (10). Moreover, two full scale upflow anaerobic
sludge blanket (UASB) reactors have been reported in the
Netherlands to treat effluents containing phenol, toluene,
benzene and benzoic acid (11).

In addition, a work on subsurface injection of wastewater

from a dimethyl terephthalate plant, which is somehow
similar to terephthalic acid wastewater, found
methanogenic activity in aquifers (12), where the waste
was highly diluted. This result suggested that
terephthalic plant effluent might be anaerobically
degraded. In fact, a previous study (13) with two UASB
reactors fed with the same wastewater utilized in this

2
work, reached 46.4% on COD removal with 2.7 days of HRT.
However, this low efficiency led us to study another type
of anaerobic reactor.

In this work, we present results from the anaerobic

treatment of a terephthalic acid plant wastewater using a
tubular fixed film reactor, in order to assess its
feasibility as an alternative to the energy consuming
aerobic process.

MATERIAL AND METHODS

Reactor design. The laboratory scale fixed film reactor

(Fig. 1) was made with a plexiglass column of 100 cm high
and 9.6 cm internal diameter, packed with PVC tubes 67 cm
high and 1.27 cm of diameter. The tubes provided 1.05 m2
of support area (specific surface area: 221 m2/m3) and a
void volume of 4.75 l. The reactor was fed in a downflow
pattern with a peristaltic pump. The influent wastewater
was kept at 6°C under mixing. Biogas was evacuated at the
top of the column and directed to a gas meter device
filled with acidified brine (pH < 4). The system was kept
at 33±2°C.

Gas meter
column

Effluent
Tubular fixed
Film reactor

6°C
Influent

Magnetic mixer

Fig.1 Schematic diagram of the experimental arrangement

3
Wastewater characteristics. The water used in this study
was obtained from a Purified Terephthalic Acid (PTA)
factory. Raw wastewater was weekly sampled at the entrance
of the activated sludge treatment system of the PTA plant,
after nutrient (N and P) addition. Table 1 shows a
characterization of the wastewater, with the main aromatic
compounds found in the mother liquor (liquid waste from
the purification unit), which is an important constituent
of the raw wastewater.

Table 1. Characteristics of the wastewater

Parameter Concentration

Raw wastewater

pH 4.5
COD 9500 mg l-1
BOD5 5500 mg l-1
TSS 2200 mg l-1
Acetic acid 1000 mg l-1
Mother Liquor*

Terephthalic acid (PTA) 2670 mg l-1

p-toluic acid 480 mg l-1
Benzoic acid 354 mg l-1
4-formylbenzoic acid 20 mg l-1

*supernatant after 10 min sedimentation

At first, the high content of fast settling suspended

solids in the raw wastewater and the low feed rate did not
permit homogeneous feeding of the reactor. In spite of a
mixing storage vessel, the plastic tubing of the
peristaltic pump clogged frequently, so it was decided to
decant the raw wastewater before feeding. This settled
wastewater was neutralized to pH 6.15 with NaHCO3 and used
to feed the reactor. A characterization of the influent,
after neutralization, is presented in Table 2.

Start-up and operation. The reactor had already developed

biofilm on the tubular support during a previous study
(14) using anaerobically adapted activated sludge from a
municipal plant as seed. The reactor was filled with tap
water and settled wastewater (1:l) and feeding started
immediately at a HRT of 10 days (organic load of 0.67 kg

4
COD m-3d-1) during three weeks in order to start sludge
acclimation. Then, the reactor was fed at two different
HRT: 5.8 days from day 21 to 87 and 3.4 days from day 88
to 164 (1.08 and 1.89 kg COD m3 d-1 of organic load.
respectively).

Table 2. Characteristics of the influent

PERIOD
HRT 5.8 d HRT 3.4 d
PARAMETER mean n s mean n s
Alk. (mg CaCO3 l-1) 2034 20 958 1967 18 315
COD total (mg l-1) 6260 23 1309 6340 20 367
COD soluble (mg l-1) 5818 14 1234 5764 16 423
N-NH4+ (mg l-1) 97 10 36.7 113 12 21.2
TS (mg l-1) 6753 11 1659 5900 13 191
TVS (mg l-1) 2977 11 1579 2913 13 97
TSS (mg l-1) 564 10 158 677 12 211
VSS (mg l-1) 308 10 137 260 12 182
n: number of samples; s: standard deviation

Analytical - methods. pH, alkalinity, total and volatile

solids (TS,TVS), total and volatile suspended solids (TSS,
VSS), total and soluble Chemical Oxygen Demand (COD),
Biochemical Oxygen Demand (BOD5) and ammonium (N-NH4+) were
measured according to Standard Methods (15) with the
following frequency: pH, alkalinity and total COD, twice a
week; soluble COD, TS, TVS, TSS, VSS and ammonium, once a
week; effluent BOD5 was analysed three times during steady
state periods. Gas production was obtained by water
displacement and methane content was measured with gas
chromatography using a thermal conductivity detector (16)

RESULTS AND DISCUSSION

Sedimentation of the wastewater resulted in a reduction of

37% of COD and 70% of TSS. Terephthalic acid has a high
specific gravity (1.5) and low solubility in water (19 mg
l-1, 25°C). As mentioned, settling prior to feeding was
done at the pH of raw influent (4.5). Kun et al. (17)
showed that precipitation of terephthalic acid begins at
pH 5.1, it is almost complete at pH 4.5 and total
separation is accomplished at pH 3.86 (pK1=3.54,
pK2=4.46). Thus, in our case, it can be expected that most

5
of PTA is in the settled solids, only a fraction remaining
in a near colloidal form (particle size around 5 µm
according to Kun et al. (17), and measured as a part of
the influent VSS (see Table 2).

The effluent total COD curve (Fig. 2) permits us to

identify transient and steady state periods. The latter
occurred during days 73 to 87 and 145 to 164,
corresponding to the two HRT applied. The former, besides
the start-up, were present in day 48 (shock load), in days
104 to 115 and 134 to 138 (no substrate feeding) and
during days 88 to 144 and 165 to 178 (change in feed
rate). Fig. 2 also presents the COD removal efficiency and
the daily gas production during the whole experimental
period.

Start-up

At the beginning, there was no gas production, although

some COD removal efficiency was achieved. The initial
diluted content of the reactor, together with the low feed
rate applied during the first three weeks, seem to be
partially responsible for this result. Another reason is
VSS retention in the reactor by physical means; in this
period, the VSS removal efficiency was 61% and the
influent had a higher VSS content(788 mg l-1).

Gas production started at day 38, and increased markedly

from day 50 to day 73, when more regular daily gas volumes
were obtained. The start-up was long, considering that the
reactor had already a well-established biofilm. Of course,
an acclimation period could be expected due to substrate
change, from acetic and propionic acids (14) to the
petrochemical influent, but the delay in gas production
suggests that, at first. the new substrate was difficult
to be degraded anaerobically.

Steady state

Day 73 was taken as the beginning of the first steady

state, corresponding to a HRT of 5.8 days and an organic
load of 1.08 kg COD m-3 d-1 (Fig. 2, day 73 to 87). The
second steady state, for a HRT of 3.4 days and a load of
1.89 kg COD m-3 d-l, was reached in day 145 and it was
maintained until day 164. Table 3 contains data from both
periods.

6
 I

10

01
O
I I IC+!
30 '
60 I ' I
90 I I '
120
I I '
150
I l I
180
I
200

TIME ( D A Y S )

Fig. 2 Evolution
Fi.g. 2 E v o l u t i o nofofinfluent
influent and effluent
and e COD,C OCOD
f f l u e n t COD, D removal
removal
efficiency ,and biogas production. o() s h o c k load.
efficiency, and biogas production.(*) shock load.

7
233
The COD removal efficiencies for both HRT are considerably
higher than those obtained in a previous study (13) with
UASB reactors and the same wastewater. In that work, the
efficiency was 46.4% in the best case, with an organic
load of 2.6 kg COD m3 d-l (HRT 2.7 days) and a gas
production of 0.47 m3m-3d-l (NTP) with 61% of CH4.

Table 3. Characterization of the steady state periods

HRT 5.8 d HRT 3.4 d

Organic load (kg COD m-3d-1) 1.08 1.89

Organic surface load (kg COD m-2d-1) 4.78 8.4
COD removal efficiency Total (%) 77.4 74.5
COD removal efficiency soluble (%) 77.5 73.0
TSS removal efficiency (%) 27.3 33
VSS removal efficiency (%) 24.0 n.d.
Gas production (Nm3m-3d-1) 0.46 0.63
CH4 yield (Nm3CH4 kg-1COD removed) 0.36 n.d.
CH4 content (%) 62 n.d.

n.d. not determined

For a more direct comparison with the results in Table 3,

data obtained with a UASB reactor operated at 3.2 days of
HRT and an organic load of 2.2 kg COD m-3d-1 were (13): COD
removal efficiency of 43.9%, gas production of 0.35m3m-3d-l
(NTP) and 63% CH4. The tubular fixed film reactor was much
better than the UASB reactor in treating this particular
wastewater. In order to explain this fact, the following
considerations must be taken into account:

First, the biomass content was very different in both

reactors: 30 g VSS for the UASB and 100.2 g VSS for the
tubular reactor: this last value was estimated by
subtracting the drained volume (2.5 1) from the void
reactor volume (4.75 l) and applying a biofilm density of
0.116 kg VSS m-2 for a 2.6 mm biofilm thickness (18).
Thus, the organic specific load applied to the UASB (0.38
kg COD kg-lVSS.d-l) was higher than the estimated load for
the tubular reactor (0.09). However, when operated at 0.10
kg COD kg-lVSS.d-l, the UASB reactor only reached 34% of
COD removal (13). Second, the biomass in a tubular reactor
is distributed in its entire volume, while in a UASB

8
reactor, the sludge is located in the bottom, forming a
sludge blanket. For an influent with toxic or inhibitory
compounds, which may be in the form of TSS, a UASB will be
more affected than a fixed film reactor. In this respect,
there is evidence that the terephthalic acid plant
wastewater may have a toxic effect on groundwater
microorganisms, even in moderate concentrations (19).
Moreover, a test in serum bottles (13) showed a total
inhibition of methanogenesis with the same wastewater
utilized in our work (0.56 g COD g-1VSS), as well as with
terephthalic acid (0.23 g g-1 VSS). As mentioned, the
influent VSS contained this acid, so a toxic concentration
could have been reached by accumulation in the sludge
b1anket.

The methane yield obtained with the tubular reactor during

the first HRT is very close to the stoichiometric value
(0.35): this suggests that the COD removed was actually
converted to methane, and no accumulation of influent
organic matter occurred. Apparently, the arrangement of
support, the distribution of biomass and the hydraulic
characteristics of the tubular reactor favoured hydrolysis
of the influent VSS and conversion of soluble substrates
to methane, reducing wastewater potential toxicity.

Transient state

An organic shock load of 15000 mg COD l-1 (2.4 times the

normal load) was accidentally applied to the reactor in
day 48. The effluent COD raised rapidly over 5000 mg l-1
(Fig. 2) and the COD removal efficiency dropped under 30%.
The fact that this occurred during the start-up period,
does not permit the proper evaluation of the shock
effects. Anyway, it seems that by day 60, the reactor had
recovered as suggested by the effluent COD curve.

The change of feed rate (day 88) from HRT of 5.8 days to
HRT of 3.4 days, corresponds to an increase of 1.8 times
the original organic load. The effluent COD raised quickly
from day 91 (Fig. 2), but feeding stopped from day 104 to
115 due to technical problems in wastewater supply; during
this period, the decrease in reactor COD was caused by
batch conditions. When feeding was restored, the effluent
COD increased sharply, but almost immediately the COD
declined toward the new steady state. Apparently, the feed
stop was not responsible for the high effluent COD in day
119, which was caused by the HRT change, and only retarded

9
the normal reactor response. This may be confirmed by a
second feed stop of five days (134 to 138), which did not
affect the reactor. This agrees with the well known
resistance of anaerobic fixed film reactors to periods
without feeding.

A second HRT change was applied in day 165 (3.4 days to

2.9 days of HRT), but only limited data were obtained
because the wastewater supply stopped. Nevertheless, the
short period of operation under the new conditions (two
weeks) permitted us to follow the effluent COD response,
which was similar to the one observed in the previous HRT
modification. Apparently, the reactor was rapidly
approaching the COD removal efficiencies obtained in the
former steady states, so the optimum HRT may well be
shorter than three days)

Anaerobic-aerobic treatment process as an alternative to

the conventional aerobic treatment

Based on our results, the treatment of terephthalic acid

plant wastewater cannot be accomplished by an anaerobic
process alone, since the suspended solids in the raw
wastewater would cause clogging and accumulation of toxic
compounds in the reactor (13). A settling operation prior
to an anaerobic tubular fixed film reactor would eliminate
this problem. A comparison between this process
arrangement and the aerobic treatment of a terephthalic
acid plant synthetic effluent (3) is shown in Table 4.

As mentioned, the HRT of 3.4 days could have been further

reduced without affecting the anaerobic COD removal
efficiency. In addition, nearly 70% of the synthetic
wastewater COD was in the form of easily biodegradable
compounds (methanol and acetic acid), which was not the
case for the actual effluent employed in this study. The
former considerations led us to think that the anaerobic
treatment of the settled influent can be realized at the
same HRTs employed in the aerobic treatment. Nevertheless,
it appears that aerobic treatment reaches higher removal
efficiencies. but it needs important energy input, due to
the high influent COD concentration. In contrast, the
anaerobic process produces a significant amount of energy
(2.8 kWh kg-lCOD fed) which may be converted to electrical
energy (around 0.55 kWh kg-1COD fed). Furthermore, if a
primary settler is added prior to the anaerobic reactor,
the VSS removal increases the global COD removal

10
efficiency to 84%. Of course, the disposal of the primary
sludge must be considered.

It should be noted that the fate of the influent TSS in

the aerobic conventional process is not known. Adsorption
of organic compounds on activated sludge has been reported
(20). In our case, there is evidence of terephthalic acid
toxicity on anaerobic bacteria, which may not be the case
for aerobic microorganisms. The primary sludge adsorption
on aerobic flocs or its sedimentation in the secondary
settler may be a way of removing it, in case it is not
biodegraded.

Table 4. Comparison between two process configurations for

terephthalic acid plant wastewater treatment

Process HRT COD removal Energy balance Ref.

configuration (d) % (kWh kg-1COD fed)

Aerobic 2 >90 (-)0.62 3

(synthetic
ww*)

Primary 3.5 84 (+)0.55 This

settling + work
anaerobic
(raw ww*)

WW*: wastewater
(-) energy consumed
(+) energy produced considering a 20% conversion effi-
ciency from chemical to electrical energy.

The anaerobic reactor effluent can be further treated with

an aerobic process, its COD/BODs ratio being 1.6. In this
configuration, the primary sludge could be fed to the
aerobic system, as it is actually done in the conventional
process, and the energy produced by the anaerobic
pretreatment should cover the needs of the aerobic
postreatment unit, resulting in a highly efficient, no
energy consuming process.

CONCLUSIONS

The downflow anaerobic tubular fixed film reactor was

successful in treating primary settled terephthalic acid
plant wastewater: at a HRT of 3.4 days, it reached a 75%

11
COD removal efficiency.

The low organic load applied (1.9 kg COD m-3d-1) shows that
the petrochemical effluent has slow anaerobic degradation
kinetics. It may be confirmed by the long acclimation
period required for reactor start-up.

The apparent inhibitory characteristic of the wastewater,

present in a previous UASB reactor study, did not affect
the tubular reactor. The different biomass distribution
patterns for both types of reactors may explain this fact.
In addition, the tubular reactor had good resistance to
shock loads and to periods without feeding.

A primary settling tank followed by an anaerobic - aerobic

treatment process may be an attractive, energy saving
option to the three stage aerobic process commonly used.

ACKNOWLEDGEMENTS

Partial financial support was obtained from the

Organization of American States (OAS) and European
Economic Community (EEC). We thank E. Castillo for
valuable technical assistance and M. del Villar for his
support and interest in this study. This work is part of
two collaboration research programs between the Institut
Français de Recherche Scientifique pour le Développement
en Cooperation (ORSTOM) and the Instituto de Ingeniería
UNAM with the Universidad Autónoma Metropolitana -
Iztapalapa.

REFERENCES

1. A.G. Bemis, J.A. Dindorf, B. Horwood and C. Samans, in

Kirk Othmer Encyclopedia of Chemical Technology, F.M.
Herman, D.F. Othmer, C.G. Overberg, G.T. Seaborg, M.
Grayson and D. Eckroth, ed., John Wiley and Sons, New
York, 1982, Vol.17, 746-758.

2. Inf. Chimie, 278, 209-215 (1986).

3. C.M. Lau, Proc. 32nd. Ind. Waste Conf., Purdue Univ.,

Ann Arbor Science, Ann Arbor, 1978, 63-74.

4. F.A. Lund and D.S. Rodriguez, J. Gen. Appl. Microbiol.,

30, 53-61 (1984).

5. S. Brugnaro and B. Polo, Inquin., 27, 65-68 (1985).

12
6. M.T. Suidan, J.T. Pfeffer and G.F. Nakhla, Anaerobic
Digestion 1988, Proc. 5th Int. Symp., E.R. Hall and P.N.
Hobson ed., Pergamon Press, 1988, 249-257.

7. L.H, Nel, J. De Haast and T.J. Britz, Biotechnol.

Lett., 6, 741-746 (1984).

8. F. Edeline, G. Lambert and H. Fatticcioni, Process

Biochem., 4, 58-60 (1986).

9. P. Vogel and J. Winter, in Poster Papers, 5th Int.

Symp. Anaerobic Digestion, A. Tilche and A. Rozzi ed.,
Monduzzi editore, Bologna, 1988, 689-694.

10. P. Sanna, Anaerobic Digestion 1988, Proc. 5th Int.

Symp., E.R. Hal1 and P.N. Hobson ed., Pergamon Press,
1988, 487-489.

11. A.J.M.L. Borghans and A. Van Driel, in Poster Papers,

5th Int. Symp. Anaerobic Digestion, A. Tilche and A. Rozzi
ed., Monduzzi editore, Bologna, 1988, 627-630.

12. J.A. Leenheer, R.L. Malcolm and W.R. White, Environ.

Sci. Technol., 10, 445-451 (1976).

13. J.P. Guyot, H. Macarie and A. Noyola, Appl. Biochem.

Biotechnol.. in press.

14. O. Monroy, A. Noyola, F. Ramirez and J.P. Guyot, in

Poster Papers, 5th Int. Symp. Anaerobic Digestion A.
Tilche and A. Rozzi ed., Monduzzi editore, Bologna, 1988,
747-751.

15. APHA, AWWA and WPCF, Standard Methods for the

Examination of Water and Wastewater, 15th edition,
Washington D.C., 1980.

16. A. Noyola, B. Capdeville and H, Roques, Water Res.,

22, 1582-1592,1988

17. X. Kun, X. Liangcai, D. Huiru, S. Jianhua and Z.

Qiubo, Water Treat., 3, 481-490 (1988).

18. L. van den Berg and C.P. Lentz, in Proc. 34th Ind.
Waste Conf., Purdue Univ., Ann Arbor Science, Ann Arbor,
1980, 319-325.

13
19. J.A. Leenheer, R.L. Malcolm and W.R. White, Physical,
Chemical and Biological Aspects of Subsurface Organic
waste injection near Wilmington, N.C. Geoloqicai Survey
Professional Paper 987, U.S. Government Printing Office,
Washington D.C., 1976, 51 pp.

20. A.C. Petrasek, I.J. Kugelman, B.M. Austern, T.A.

Pressley, L.A. Winslow and R.H. Wise, J. Water Pollut.
Control Fed., 55. 1286-1296 (1983)

14