Biotreatment of Hydrate-Inhibitor-Containing

See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/277354428
Biotreatment of Hydrate-Inhibitor-Containing Produced Waters at Low pH
Article in SPE Journal · May 2015
DOI: 10.2118/174551-PA
CITATIONS
2
READS
133
6 authors, including:
Some of the authors of this publication are also working on these related projects:
Membrane Distillation Desalination View project
MFI Cranfield View project
Arnold Janson
ConocoPhillips
30 PUBLICATIONS 375 CITATIONS
SEE PROFILE
Ana Santos
ConocoPhillips
18 PUBLICATIONS 324 CITATIONS
SEE PROFILE
Altaf Hussain
26 PUBLICATIONS 286 CITATIONS
SEE PROFILE
Simon jon Judd
Cranfield University
230 PUBLICATIONS 9,471 CITATIONS
SEE PROFILE
All content following this page was uploaded by Ana Santos on 06 September 2016.
The user has requested enhancement of the downloaded file.
Biotreatment of Hydrate-InhibitorContaining Produced Waters at Low pH
Arnold Janson, Ana Santos, and Altaf Hussain, ConocoPhillips Global Water Sustainability Center; Simon Judd,
Qatar University/Cranfield University; Ana Soares, Cranfield University; and Samer Adham, ConocoPhillips Global
Water Sustainability Center
Summary
With proper treatment to remove organics and inorganics, one can
use the produced water (PW) generated during oil-and-gas extraction as process water. Biotreatment is generally regarded as the
most cost-effective method for organics removal, and although
widely used in industrial wastewater treatment, PW biotreatment
installations are limited.
This paper follows up to an earlier paper published in the SPE
Journal (Janson et al. 2014). Although the earlier paper assessed
the biotreatability of PW from a Qatari gas field from the summer
season, this paper focuses on assessing the biotreatability of PW
during the winter season [i.e., containing the thermodynamic
hydrate inhibitor monoethylene glycol (MEG) and a kinetic
hydrate inhibitor (KHI)]. Tests were conducted in batch and continuous reactors under aerobic mixed-culture conditions without
pH control during 31 weeks.
The results indicated that one could remove >80% of the chemical oxygen demand (COD) and total organic carbon (TOC)
through biological treatment of PW with 1.5% MEG added. In contrast, biotreatment can remove only 43% of COD and TOC present in PW when 1.5% KHI was added as a hydrate inhibitor; 2-
butoxyethanol, a solvent in KHI, is extremely biodegradable; it was
reduced in concentration from >5000 to <10 mg/L by biotreatment; the KHI polymer though was only partially biodegradable.
Cloudpoint tests conducted on PW with 1.5% KHI added showed
only an 8C increase in cloudpoint temperature (from 35 to 43C).
The target cloudpoint temperature of >60C was not achieved.
Although the feed to the reactors (PW with either KHI or
MEG) was at pH 4.5, the reactors stabilized at a pH of 2.6, considered extremely acidic for aerobic bioactivity. The successful operation of an aerobic biological process for an extended period of
time at a pH of 2.6 was unexpected, and published reports of bioactivity at that pH are limited. After extensive analytical tests, it
was concluded that the pH decrease was caused by the production
of an inorganic acid. A mechanism by which hydrochloric acid
could be produced biologically was proposed; however, further
research in this area by the academic community is recommended.
Introduction
Produced water (PW) arising from oil-and-gas (O&G) exploration
represents the largest-volume byproduct from petroleum production. Even though generally treated for reinjection or discharge,
there is increasing interest in its reuse as process water for O&Gindustry operations. This then requires removal of both “natural”
and “field” chemicals, the latter relating to those reagents added
at the wellhead to expedite operation including scale-, corrosion-,
and hydrate-inhibiting chemicals.
Hydrates are manifested as ice-like, methane-containing crystals that can block pipelines and process equipment, leading to
significant downtime. Hydrate inhibitor is added during the lowtemperature winter months at natural-gas wells in Qatar to prevent
the formation of methane hydrates in the pipelines between the
offshore platform and the onshore processing facilities. There are
two primary classifications of hydrate inhibitors:
• Kinetic hydrate inhibitor (KHI), a proprietary polymer
offered by chemical suppliers
• Thermodynamic hydrate inhibitor [e.g., monoethylene glycol (MEG)]
KHI is added at a concentration between 1 and 2%, and MEG
is added at concentrations of >15%.
This paper focuses on assessing the biotreatability of PW containing hydrate inhibitors with the activated-sludge process without pH control. To ensure stable composition during the course of
the study, approximately 1200 L of PW was collected from a
Qatar gas field during the summer season (i.e., without hydrateinhibitor dosing). After collection and stripping to remove H2S,
the PW was stored in a large plastic tank at room temperature.
Eight months of biotreatability testing on this water was reported
previously in this journal (Janson et al. 2014). The tests reported
in this paper were conducted after the summer-season tests on the
same PW sample with KHI and MEG added at the 1.5% concentration, thereby allowing assessment of their comparative biotreatability when arising in winter-season PW.
An unexpected result of this study was the highly acidic pH
(2.6) of the reactor under long-term steady-state conditions. Aerobic biological treatment of PW by activated sludge was previously
reported for classical activated sludge (Tellez et al. 2005),
sequencing batch reactors (Freire et al. 2001; Pendashteh et al.
2010), and, most recently, for membrane bioreactors (Pendashteh
et al. 2012; Kose et al. 2012; Sharghi et al. 2013). However, very
few reports were found of aerobic biotreatment under the unusual
conditions of extremely low pH encountered in the current study.
A survey of the literature in this area reveals low-pH (pH <3) biochemistry to be largely associated with acidithiobacillus sulfide:-
sulfate transformations (Dopson and Johnson 2012), which can
take place at pH values as low as zero (Macalady et al. 2007).
Biotreatment processes for classical aerobic carbonaceous removal, however, appear to have a lower pH limit of approximately 5.5 imposed by nitrification (Metcalf and Eddy 2004).
Materials and Methods
Produced Water (PW) and Reagents. Approximately 1200 L
of PW was collected from a Qatari gas field during summer season and stored in a 1400-L tank at room temperature. This water
was aerated to strip it of hydrogen sulfide, reducing the initial
chemical oxygen demand (COD) of 1600 to 1100 mg/L. The
PW was moderately saline (3200 mg/L of NaCl, 7200-lS/cm conductivity), with a total organic carbon (TOC) of 491 mg/L and a
total N content of 34 mg/L. Analyses of this PW showed it to contain oil and grease and salinity at concentrations within the range
of those normally encountered for such waters (Fakhru’l-Razi
et al. 2009). In addition to small amounts of hydrocarbon and formation constituents, the PW contained field chemicals relating to
corrosion and scaling mitigation, but contained no hydrate inhibitors because it was collected during the summer season when the
hydrate-formation potential is very low.
To represent the PW composition during winter, kinetic
hydrate inhibitor (KHI) and monoethylene glycol (MEG) were
added separately, each at a concentration of 1.5%. The KHI used
was a commercial product consisting of a blend of a proprietary
Copyright VC 2015 Society of Petroleum Engineers
Original SPE manuscript received for review 26 October 2014. Revised manuscript received
for review 20 January 2015. Paper (SPE 174551) peer approved 23 March 2015.
2015 SPE Journal 1
polymer and a solvent, 2-butoxyethanol (2-BE). The 2-BE is 30 to
60% of the commercial KHI product. After the KHI addition, the
feedwater COD increased from 1100 to 31 500 mg/L, indicating
that the KHI contributed >95% of the total COD. The addition of
MEG at the same percentage concentration increased the feed
COD to 21 800 mg/L. Nitrogen, phosphorous, and potassium were
added as nutrients to the winter PW as ammonium chloride and
potassium phosphate monobasic in amounts in excess of that
required for bio-oxidation. Nitrogen compounds were determined
as NH4-N, NO2-N, NO3-N, and TN, and phosphorus measured as
phosphate (PO43–). The compositions of the winter-season PW
with hydrate inhibitors and nutrients are presented in Table 1.
The biomass used to seed the KHI batch reactor was from earlier extended experiments assessing the biotreatability of PW under
summer-season operating conditions (Janson et al. 2014). This biomass was subsequently acclimated to the KHI feed during 6 weeks.
For the MEG-fed trials, the reactor was seeded from the KHI batch
reactor supplemented with biomass from earlier summer trials and
with dried biomass used in standard biochemical oxygen demand
tests. Because bioactivity was quickly evident, the biomass acclimation period on PW containing MEG was only 10 days.
Bioreactor Apparatus. Tests were conducted in batch reactors
and a continuous membrane bioreactor (MBR). The MBR featured a hollow-fiber immersed ultrafiltration membrane in a bioreactor with a 1-L working volume. Details on the MBR design
are in Janson et al. (2014). As noted in the Organics Removal subsection, the higher aeration intensity in the MBR seemed to cause
a decrease in bioactivity, and all results reported in this paper pertain to the batch bioreactors.
The laboratory-scale batch reactors consisted of 4-L polyethylene vessels with a working volume of 1 L (Fig. 1). Each reactor
was aerated with a ceramic frit connected to an air pump and continuously magnetically stirred. Probes for sensing dissolved oxygen (Hach Lange DO meter), pH (Thermo Scientific), and
conductivity (Thermo Scientific) were immersed in the batch reactor for the daily measurement of these parameters. The aeration
intensity in the reactor was 0.0010 m3/s of air per m2 of reactorsurface area. In contrast, the aeration intensity in the MBR was
75X greater, at 0.076 m3/s of air per m2 of reactor-surface area.
The ambient temperature in the laboratory was typically 20C.
For the long-term steady-state tests, a 10-day hydraulic retention time and a 40-day solids residence time (Table 2) were maintained in both batch reactors to sustain a predicted food/
microorganism (F/M) ratio of 0.5. This was considered to be the
maximum F/M ratio viable on the basis of literature values (Metcalf and Eddy 2004). To compensate for evaporative losses,
deionized (DI) water was added to the batch reactor at a rate of
30 cm3/d to maintain the total volume at 1 L. Treated effluent
was generated by centrifuging approximately 200 cm3 of batchreactor contents for 20 minutes at 3000 g and collecting 100 cm3
of supernatant; the remaining supernatant and solids were
returned to the batch reactor. Feedwater was added twice daily in
50-cm3 aliquots to maintain the process. For comparison, tests
were subsequently conducted under neutral pH conditions; the
KHI-fed batch reactor was manually adjusted daily to a set point
of 7.0 to 7.2 with 1-M sodium hydroxide, with the pH generally
kept between 5.9 and 7.2. The MEG batch reactor was only operated at a pH of 2.6 to determine if another carbon source could be
degraded similarly at that low pH or if KHI chemistry were a factor. MEG is known to be biodegradable at neutral pH conditions.
Monitoring and Analysis. Biomass dissolved-oxygen (DO) concentration, oxygen-uptake rate (OUR), temperature, and pH were
monitored daily, along with feed and treated effluent pH and conductivity. Samples of feed and treated effluent were collected
three times a week for COD and TOC assays. During the final 2
weeks of steady-state tests, a 50-cm3 sample of biomass was centrifuged, and the solids were subsequently dried and ashed to
determine total and suspended volatile solids (TSS, VSS). Also,
during these 2 weeks, additional samples of treated effluent and
feedwater were taken for total nitrogen and ion analysis (including nitrite and nitrate).
The OUR was measured in-situ by halting batch-reactor aeration, and with mixing still in effect, recording the DO concentration every minute for 5 minutes. The OUR, expressed as mg O2/
Lhr, was then calculated from the slope of DO concentration vs.
time graph. If the DO concentration reached a value of less than
1.3 mg/L, the OUR test was considered complete because the
Table 1—Winter-season produced-water composition with nutrients
added.
Air pump Magnetic stirrer
1-L working volume in a 4-L vessel
DO, pH, Conductivity,
Temperature meter
Fig. 1—Laboratory-scale batch-reactor apparatus.
2 2015 SPE Journal
organisms were observed to be oxygen-limited. The DO concentration and OUR were monitored during 8 hours after feeding to
determine the time required for the DO concentration to return to
its original value, indicating the length of time for the organisms
to consume the feed added.
COD was determined on the basis of standard dichromatedigestion methodology (APHA 2005) with a Hach spectrophotometer. Cation and anions, including organic acids, NH4-N, NO2-N,
and NO3-N, were analyzed by ion chromatography (IC, Dionex),
on the basis of IonPac CS 12A and AS 19 columns with EGC III
MSA/KOH eluent generator coupled with ion detection by conductivity. Individual elements were determined by inductively coupled
plasma optical emission spectroscopy (ICP-OES) with a Thermo
Scientific iCAP 6000 Series instrument. TOC and TN were determined with a Shimadzu TOC analyzer by use of vendor-provided
protocols, with TOC calculated by difference between the total and
inorganic carbon. The TSS and VSS were determined by centrifugation of 50-cm3 biomass samples for 20 minutes at 2,500 rev/min
followed by gravimetric determination (105C for 6 hours, ignition
at 525C for 25 minutes) of the solid pellet fraction and total sample; the VSS was determined by difference.
The KHI content in the samples was measured on the basis of
a modified standard iodine complexation (iodometric) method;
tri-iodide (KI3) complexed with the ethylene oxide chain (functional groups) of KHI resulted in a color shift that varied directly
with the KHI concentration. The color shift was measured with a
ultraviolet–visible spectrometer at 500 nm after 10 minutes of
reaction time. The MEG content was measured indirectly through
TOC and COD measurements.
The 2-BE was measured by gas chromatography (GC)
equipped with mass spectrometer and flame-ionization detector
(GC-MS/FID, Agilent 7890). Standard solutions of 2-BE (100 to
200 mg/L) were prepared in DI water. Both feed and effluent samples were filtered (0.45-lm nylon filter) and diluted 100 times
before analysis by direct aqueous injection (0.2 lL, split mode) at
250C into a polar GC column (HP-Innowax, 19091N-133). The
oven temperature was programmed at 40C for 5 minutes, raised
from 40 to 250C at 30C/minute, and held at 250C for 8
minutes. Helium was the carrier gas.
Efforts were made toward identifying any biochemically generated organic acid responsible for lowering the pH. These analyses were conducted with a Thermo LTQ XL ion trap MS with
direct sample injection. The samples were analyzed by electrospray ionization MS (ESI-MS) in both positive and negative ion
modes, scanning a mass range of 150 to 2,000 Da. Volatile organic acids were also screened with the same GC-MS/FID method
used to measure 2-BE. Ion-exclusion ion chromatography (Dionex, ICS-900 series) with an ICS-AS1 column was used to determine acidic anions (glycolate, formate, acetate, propionate, and
butyrate). The chromatography was carried in ion-suppression
mode with tetrabutylammonium hydroxide (5 mM) as regenerant
and 1-heptafluorobutyric acid (0.4 mM) as eluent.
A series of separate abiotic tests was conducted to assess the
impact of stripping on TOC removal. In these tests, various concentrations of 2-BE and MEG were added to either 1 L of DI
water or PW, and the solution aerated and mixed for 24 hours
under the same conditions as the batch-reactor trials. Samples for
TOC and COD analysis were taken at t ¼ 0, 6, and 24 hours.
Results and Discussion
Organics Removal. After seeding, the biomass was acclimated
in a batch reactor to produced water (PW) with kinetic hydrate inhibitor (KHI) added. In the beginning, the bioreactor pH was manually adjusted daily to 5.5 6 0.1 pH units by adding caustic. This
pH adjustment was considered necessary because nitrification generally requires a pH of >5.5. Within a few days of operation
though, it became apparent that active biological activity was
occurring and acids were being generated that were capable of
decreasing the feed pH from 4.5 to less than 3.0. Because oxygenuptake-rate (OUR) and chemical-oxygen-demand (COD) measurements indicated a high rate of biological activity at a pH <3, manual pH adjustment was suspended, and the batch reactor was
operated without pH adjustment.
After 6 weeks of acclimation to the PW with gradually
increasing amounts of KHI in a batch reactor, the biomass was
transferred to a membrane bioreactor (MBR) to assess bioactivity
in a continuous mode. During a period of 1 to 2 weeks, the biological activity gradually decreased. This trend was indicated
through: (i) a decrease in the OUR; (ii) decreases in total-organiccarbon (TOC) and COD removal rates; and (iii) an increase in pH
to >3.5. It was decided to return the biomass back to the batch reactor in the hope of recovering bioactivity. Within a week, the
bioactivity was restored (high OUR, pH >3), and a second
attempt was made at operation in continuous mode. Again, the
biomass activity decreased. The speculation was that the bioactivity decreased because the significantly higher aeration intensity of
the MBR compared with the batch reactor (0.076 m3/m2s vs.
0.0010 m3/m2s) resulted in physical shearing of the fungus. It
was decided to conduct 14 weeks of steady-state KHI tests exclusively in a batch reactor.
Analytical and performance data were determined as mean
values for the final 2 weeks of operation under steady-state conditions. For the KHI-fed batch-reactor system, data are presented
for both unadjusted pH (2.6) and the adjusted pH (7.2). For the
monoethylene glycol (MEG) -fed system, testing was only conducted at the unadjusted pH (2.6) condition.
Results indicate that TOC and COD were removed at rates of 42
to 44% in the KHI-fed batch reactor, independent of pH (Table 3).
Corresponding TOC and COD removals for an MEG-fed reactor
were 85 and 81%, respectively. The significantly higher TOC and
COD removals for the PW with MEG added are consistent with
expectations because MEG is generally considered to be highly
biodegradable. Unlike the KHI biomass for which the solids concentration was stable, the MEG biomass also experienced a
reduction of 37% in mixed liquor volatile suspended solids
(MLVSS) (from 6 to 4 g/L) during the course of the study. One
should note that the MLVSS measurements are not considered
very accurate because of the long hydraulic retention time (HRT)
and small sample volumes (50 cm3) that were used in the tests.
Because of the inherent MLVSS-measurement inaccuracy, one
could not estimate accurately the yield.
The 2-butoxyethanol (2-BE) concentration data indicated that
although the feed concentration was in excess of 5000 mg/L, the
concentration of the solvent in the effluent was <10 mg/L. It is
noted that a significant portion of the COD removed can be
directly attributed to the bio-oxidation of 2-BE, but there is evidence that some KHI polymer was also oxidized.
Although KHI measurements were conducted, the actual
results are not reported because a ConocoPhillips-supplier confidentiality agreement is in effect.
Even though ammonia and phosphorus were always detected
in the effluent, neither nitrate nor nitrite was detected in the
treated water at a detection limit of 1 mg/L. This result and the
fact that nitrification and denitrification are not expected to occur
at a pH of 2.6 would indicate that the ammonia removed from
the feed was used in the production of biomass, rather than converted to nitrate or nitrogen gas. The average reduction in ammonia (NH4þ) across the batch reactor was 715 mg/L as
ammonium for the KHI-fed system (589 mg/L as N). With the
assumption of a 5:1 ratio of carbon to nitrogen (Metcalf and
Eddy 2004), this equates to a TOC reduction of 2940 mg/L,
lower but reasonably consistent with observed TOC removals of
4100 to 4300 mg/L.
Solids residence time, days
Table 2—Operating conditions.
2015 SPE Journal 3
Acid Production and Batch-Reactor pH of 2.6. The experimental results demonstrated that the biological oxidation of 2-BE (carrier solvent in commercial KHI product) or MEG resulted in
unexpected acid production and a decrease in reactor pH to 2.6. In
the case of the 2-BE, this activity was demonstrated continuously
in various batch and membrane bioreactor tests during 31 weeks.
The MEG activity was for comparison purposes, and tests were
conducted only during a 10-week period.
Analyses of the treated effluent were conducted to determine
whether an organic acid was ultimately accountable for the pH of
2.6. Even though ion chromatography analyses showed a minor
amount of acetate in the PW feed (350 mg/L) and trace amounts
of other organic acids, the treated effluent showed <1 mg/L of
low-molecular-weight carboxylic acids. Mass-spectrometry (MS)
and gas-chromatography/MS (GC/MS) analyses revealed no detectable amounts of organic acids present in the effluent (and thus,
by implication, the reactor). Although the biological formation of
2-butoxyacetic acid from 2-BE appears possible and could have
contributed to the pH shift observed, there was no evidence of this
chemical in the system. Also, the TOC and COD removals would
indicate that the 2-BE was virtually completely removed and not
simply partially oxidized to 2-butoxyacetic acid. It was concluded
that the acidity is inorganic in nature.
The mechanism by which inorganic acids can be produced biologically is unknown, and literature reports were not found to
explain the observed phenomenon. In one reference, Aspergillis
niger is reported to have obtained a pH of 2.4 two days after inoculation of a 10% (v/v) activated sludge from municipal wastewater treatment. The pH is taken to be a result of the production
of “acidic metabolites” (Mannan and Fakhru’l-Razi 2005).
The following mechanism is hypothesized to explain the biological production of an inorganic acid [[hydrochloric acid (HCl)]:
1. Organisms absorb an organic compound into the cell and oxidize it, producing an organic acid (carboxylate anionþ Hþ).
2. Nitrogen, a nutrient necessary for protein synthesis and provided as NH4Cl, is absorbed into the cell, and simultaneously, the Hþ ion generated by organics oxidation is
excreted from the cell (exchanged) to maintain electrical
neutrality.
3. The carboxylate anion and nitrogen are used in biomass
synthesis, and the resulting byproduct is HCl, an inorganic
acid capable of reducing the pH to 2.6.
The three steps are represented schematically in Fig. 2.
There is speculation that bioactivity at pH <3 has not been frequently observed or studied for the following reasons:
a. Anerobic oxidation is generally considered more cost-effective for high-strength COD wastewaters so there is less likelihood for researchers to consider aerobic oxidation for a
30 000 mg/L COD feed.
b. The long 10-day (240-hour) HRT meant any acids produced
would likely remain in the system for a much longer time
than conventional bioreactors that would be operated at typically 6- to 18-hour HRT.
c. Researchers would typically provide bioreactor pH control
to near-neutral pH conditions because that is considered
“standard practice” for aerobic biological treatment.
d. The feed pH of 4.5 is low and therefore contains no alkalinity to act as a buffer.
e. Because of the high COD, the ammonia consumption was
very high (715 mg/L).
f. The high COD and ammonia consumption mean the
amounts of Hþ and Cl– produced would be sufficient to
result in the observed pH decrease.
g. Because the organisms failed to survive in an MBR, it
would indicate that they are shear-sensitive, and aeration intensity needs to be considered in bioreactor operation.
h. The neutral pH typical of wastewater treatment favors the
growth of bacteria rather than filamentous fungi.
i. Filamentous fungi are not desirable in wastewater-treatment
systems because their growth represents the most common
operational problem in activated-sludge processes (Metcalf
and Eddy 2004).
Table 3—Analytical and performance summary, mean (6 standard deviation).
Step 1: Feed oxidized to organic acid Step 2: Ammonia exchanged with H+ Step 3: HCl as byproduct

Fig. 2—Proposed mechanism for biological production of HCl.
4 2015 SPE Journal
Under this hypothesis, the amount of acid produced is in direct
proportion to the amount of ammonia consumed, and both the
acid production and resulting pH can be estimated:
• Ammonia consumption: 715 mg/L
• Moles of ammonia consumed: 0.042 mol/L
• Moles of Hþ and Cl– produced: 0.042 mol/L
• Resulting pH: 1.4
Although this is significantly lower than experimental results,
one should note that on two separate occasions, the bioreactor pH
dropped below the minimum range for the pH meter (i.e., <2.0).
There is also the likelihood that some buffering of the pH is provided by other biological metabolites. It is speculated that this
mechanism is not unique to 2-BE and MEG. It could perhaps
occur with other organics that are provided to fungi at high concentrations in a feed with no alkalinity and with ammonium chloride as a source of nitrogen.
OUR. OUR was the key parameter used to monitor the activity of
the biological population; during the course of the 31 weeks of
KHI tests, the OUR and pH were each measured an estimated 150
times. Although the pH was generally stable throughout those 31
weeks (range 2.0 to 2.8), the final 14 weeks were considered the
steady-state period, and OUR and pH plots are provided (Fig. 3a).
After Week 11, the reactor pH was increased through daily manual
adjustment gradually reaching a pH of 7 in Week 13. The OURs
under both pH regimes were comparable, indicating that the low
pH was not a factor in limiting biological oxidation of KHI.
For MEG, the pH was stable during the 10 weeks of testing,
and the OUR varied between 12 and 90 mg O2/Lhr (Fig. 3b). The
gradual decrease in OUR was likely a result of the 37% decrease
in MLVSS noted previously.
To assess if the biomass had consumed all the feed provided
the previous day, the OUR was measured before feeding. Similarly, to assess if the biomass were feed-limited, the OUR was
again measured soon after feeding. OUR measurements before
feeding were consistently <15 mg O2/Lhr, whereas after feeding,
the OURs averaged 114 and 102 mg O2/Lhr at pH values of 2.6
and 7.2, respectively, confirming biological activity is independent of the batch-reactor pH.
On two separate occasions, a comprehensive OUR profile was
generated by measuring the OUR both before and after feeding
and at various intervals thereafter (Fig. 4a). OUR reached a maximum of 132 mg O2/Lhr after 120 minutes and gradually
decreased thereafter when the metabolic activity became feedlimited. After 8 hours, the OUR returned to “prefeeding” levels (6
mg O2/Lhr), indicating that virtually all the biodegradable carbon
was consumed. The pattern was repeated on dosing with another
50 cm3 of feedwater.
Integration of the OUR vs. time curve to determine the mass
of oxygen consumed by the organisms in oxidizing the feedwater
organics revealed the total oxygen consumed to be 569 mg of O2
during the 8-hour test. If oxygen used for cell maintenance is
ignored, this oxygen consumption should equate to the amount of
COD removed by the microorganisms. The 50 cm3 of feedwater,
at a COD of 30 200 mg/L, contributed a total of 1510 mg of COD
of which 42% was removed on the basis of influent and effluent
COD concentrations. This 42% equates to 634 mg/L of COD
removed and agrees reasonably well with the 569 mg of O2 consumed by the organisms during the same interval.
Similar OUR results for the MEG-fed system (Fig. 4b) indicated biodegradation occurring at the reactor pH of 2.6, although
the decrease in the dissolved-oxygen concentration on adding the
feedwater was not as significant as that observed with KHI. OUR
measurements 5 minutes after feeding averaged 58 mg O2/Lhr.
Measurement of OUR during an 8-hour period (Fig. 4b), in the
same way as for the KHI tests, revealed a short-term spike in
140 8
7 6 5 4 3 2 1 0
OUR
pH
OUR
pH
pH
8 7 6 5 4 3 2 1 0
pH
With pH
adjustment
120
100
80
OUR (mg O2/L.hr)
60
Without pH adjustment
40
20
0
0 2 4 6 8 10
Week
(a)
12 14 16
140
120
100
80
OUR (mg O2/L.hr)
60
40
20
0
0 2 4 6 8 10
Week
(b)
12
Fig. 3—The pH and OUR weekly average for KHI during 14 weeks (a) and for MEG during 10 weeks (b).
140
160 70
60
50
40
30
20
10
0
0 1 2 3 4 5 6 7 8 9
Time (minutes)
(a)
0 1 2 3 4 5 6 7 8 9
Time (hours)
(b)
120
100
80
OUR (mg O2/L min)
OUR (mg O2/L.hr)
60
40
20
0
Before
feeding
Before
feeding
After
feeding
After
feeding After
feeding
After
feeding
Area under curve represents total
oxygen consumed ≈ COD removed
Fig. 4—OUR profile measured during 8 hours for batch reactor fed with (a) KHI and (b) MEG.
2015 SPE Journal 5
OUR after feeding but with a return of the OUR to prefeeding levels within 15 minutes. Contrary to the trend observed for the KHIfed batch reactor, the OUR remained stable, indicating that bioactivity was biomass-limited rather than feed-limited during this
8-hour period.
Microbiology. Because carbonaceous removal by bacteria would
not be expected to take place at the low pH levels recorded in this
study, attempts were made to identify the microorganism responsible for the observed TOC and COD removals and OURs. Microscopic examination revealed an abundance of filamentous
microorganisms (Fig. 5a) closely resembling that reported for the
fungi Geotrichum (Fig. 5b). A study of two full-scale MBR activated sludges revealed this organism to be the most abundant fungi
during the course of the 9-month study for both aerobic and anoxic
sludge samples (Awad and Kraume 2011). Moreover, its formation
appears to be favored at high COD loading rates in reactors fed
with acetate (Matos et al. 2012), as was the case in this study.
Although there were no reports of the application of Geotrichum, or fungi generally, at very low pH for carbonaceous removal in aerobic bioreactors, it was demonstrated that sludge
inoculated with the filamentous fungus Aspergillus niger can
result in >90% COD removal through fermentation of US wastewater sludge (Mannan and Fakhru’l-Razi 2005). In another study,
a minimum pH of 2.2 to 2.3 was reached after 2 days, which was
previously attributed to the excretion of acidic metabolites (Fakhru’l-Razi et al. 2002). There were further reports of isolated Aspergillus niger strains growing at pH levels of 1 to 2.5 (Isobe
et al. 2006; Hanapi et al. 2011). Although no comparable reports
at low pH appear to exist for Geotrichum, it seems plausible, at
least under the batch and extended HRT conditions used in the
current study, that a culture developed at pH values as low as 2.6
can be used for biotreatment of the wastewaters tested in the
current study.
Cloudpoint. The cloudpoint is the temperature at which turbidity
is first observed when the temperature of the PW (containing
hydrate inhibitor) is increased, as a result of the formation of a
precipitate. One of the objectives of the biotreatment of the PW
water was to raise the cloudpoint to 60C or higher to reduce the
risk of precipitation and plugging of the reservoir when the PW is
reinjected into the formation. In the KHI-dosed feedwater, the
cloudpoint was 35C; biotreatment increased this only slightly to
42 to 45C. Thus, although significant COD was removed through
biotreatment, these results suggest that there was limited biodegradation of the KHI polymer. This is consistent with biotreatment removing primarily the 2-BE without oxidizing the largermolecular-weight polymer.
Stripping. Abiotic stripping tests were conducted at various concentrations of 2-BE and MEG, with CODs measured at various
intervals up to 24 hours. For 2-BE, COD losses from stripping
were estimated at 4 to 6%, whereas for MEG, no significant COD
was lost by stripping.
Conclusions
After extensive tests to assess the biotreatability of winter-season
produced water (PW) (i.e., with hydrate inhibitors added) in
Qatar, it was concluded that
• For PW with 1.5% monoethylene glycol (MEG) added, biological treatment is capable of removing >80% of the chemical oxygen demand (COD) and total organic carbon (TOC).
• In contrast, only 43% of COD and TOC present in PW
with 1.5% kinetic hydrate inhibitor (KHI) can be removed
biologically.
• The 2-butoxyethanol, the solvent in KHI, is extremely biodegradable; it was reduced in concentration from >5000 mg/L
to <10 mg/L by biotreatment; the KHI polymer though was
only partially biodegradable.
• Cloudpoint tests conducted on PW with 1.5% KHI added
showed only an 8C increase in cloudpoint temperature
(from 35 to 43C). The target cloudpoint temperature of
>60C was not achieved.
• Without pH control, the pH of a reactor treating PW with either KHI or MEG stabilized at a pH of 2.6, considered
extremely acidic for biological organisms.
• Very active biological activity was sustainable at a pH of
2.6; the oxygen-uptake rates measured 5 minutes after feeding were 114 mg O2/Lhr for KHI and 58 mg O2/Lhr for
MEG.
• For the PW with KHI added, adjustment of the reactor pH to
7 did not result in higher biological activity (as measured
by OUR), nor did the TOC or COD removals change.
The successful operation of an aerobic biological process for
PW treatment during an extended time at a pH of 2.6 was unexpected, and reports of active bioactivity at that pH are limited. After extensive analytical tests, it was concluded that the pH
decrease was because of the production of an inorganic acid. A
mechanism by which hydrochloric acid could be produced biologically was proposed; however, further research in this area by the
academic community is recommended.
Acknowledgments
This publication was made possible by NPRP Grant # 5-573-1-
102 from the Qatar National Research Fund (a member of Qatar
Foundation). The statements made herein are solely the responsibility of the authors.
The authors gratefully acknowledge the expertise provided by
Samir Gharfeh, Isik Turkmen, Nabin Upadhyah, and Eman AlShamari for their analytical skills and, in particular, for their help
in trying to identify the compounds produced by the organisms
that resulted in the low batch-reactor pH. The authors would also
(a) (b)
Fig. 5—Optical-microscope observation at 600X magnification: (a) current study, KHI-fed biomass with 40 days solids residence
time, (b) Geotrichum candidum sample (Thunderhouse Instruments website, accessed in 2014).
6 2015 SPE Journal
like to acknowledge the assistance of Zaid Chowdhury in the editing of this manuscript.
References
APHA. 2005. American Public Health Association, American Water
Works Association (AWWA) and Water Environment Federation
(WEF): Standard Methods for the Examination of Water and Wastewater, twenty-first edition.
Awad, M. F. and Kraume, M. 2011. Fungal Diversity in Activated Sludge
From Membrane Bioreactors in Berlin. Can J Microbiol. 57 (8):
693–698. http://dx.doi.org/10.1139/w11-056.
Dopson, M. and Johnson, D. B. 2012. Biodiversity, Metabolism and
Applications of Acidophilic Sulfur-Metabolizing Microorganisms.
Environ. Microbiol. 14 (10): 2620–2631. http://dx.doi.org/10.1111/
j.1462-2920.02749.x.
Fakhru’l-Razi, A., Alam, M. Z., Idris, A. et al. 2002. Domestic Wastewater Biosolids Accumulation by Liquid State Bioconversion Process
for Rapid Composting. J. Environ. Sci. Health, Part A: Toxic/Hazard.
Subst. Environ. Eng. 37 (8): 1533–1543. http://dx.doi.org/10.1081/
ESE-120013275.
Fakhru’l-Razi, A., Pendashteh, A., Abdullah, L. C. et al. 2009. Review of
Technologies for Oil and Gas Produced Water Treatment. J. Hazardous Materials 170: 530–551. http://dx.doi.org/10.1016/j.jhazmat.2009.
05.044.
Freire, D. D. C., Cammarota, M. C., and Sant’Anna, G. L. 2001. Biological Treatment of Oil Field Wastewater in a Sequencing Batch Reactor.
Environ. Technol. 22: 1125–1135. http://dx.doi.org/10.1080/0959333
2208618203.
Hanapi, N. H. M., Alam, M. Z., and Karim, M. I. A. 2011. Optimization
of Process Parameters for the Production of Citric Acid Using Sewage
Sludge as Major Fermentation Media. Afr. J. Biotechnol. 10 (81):
18807–18810. http://dx.doi.org/10.5897/AJB11.2757.
Isobe, K., Inoue, I., Takamatsu, Y. et al. 2006. Production of Catalase by
Fungi Growing at Low pH and High Temperature. J. Biosci. & Bioeng.
101 (1): 73–76. http://dx.doi.org/10.1263/jbb.101.73.
Janson, A., Santos, A., Katebah, M. et al. 2014. Assessing the Biotreatability of Produced Water From a Qatari Gas Field. SPE J. SPE-173188-
PA. http://dx.doi.org/10.2118/SPE-173188-PA (in press; posted 3 December 2014).
Kose, B., Ozgun, H., Ersahin, M. E. et al. 2012. Performance Evaluation
of a Submerged Membrane Bioreactor for the Treatment of Brackish
Oil and Natural Gas Field Produced Water. Desalination 285:
295–300. http://dx.doi.org/10.1016/j.desal.2011.10.016.
Macalady, J. L., Jones, D. S., and Lyon, E. H. 2007. Extremely Acidic,
Pendulous Cave Wall Biofilms From the Frasassi Cave System, Italy.
Environ. Microbiol. 9: 1402–1414. http://dx.doi.org/10.1111/j.1462-
2920.2007.01256.x.
Mannan, S., Fakhru’l-Razi, A. and Alam, M. Z. 2005. Use of Fungi To
Improve Bioconversion of Activated Sludge. Water Res. 39 (13):
2935–2943. http://dx.doi.org/10.1016/j.watres.2005.04.074.
Matos, M., Pereira, M. A., Nicolau, A. et al. 2012. Influence of the Organic Loading Rate on the Growth of Galactomyces Geotrichum in
Activated Sludge. J. Environ. Sci. Health, Part A: Toxic/Hazard.
Subst. Environ. Eng. 47 (4): 565–569. http://dx.doi.org/10.1080/
10934529.2012.650563.
Metcalf and Eddy. 2004. Wastewater Engineering: Treatment and Reuse,
fourth edition, New York: McGraw-Hill.
Pendashteh, A. R., Abdullaha, L. C., Fakhru’l-Razi, A. et al. 2012. Evaluation of Membrane Bioreactor for Hypersaline Oily Wastewater Treatment. Process Safety & Environ. Protection 90: 4555. http://
dx.doi.org/10.1016/j.psep.2011.07.006.
Pendashteh, A. R., Fakhru’l-Razi, A., Chuah, T. G. et al. 2010. Biological
Treatment of Produced Water in a Sequencing Batch Reactor by a
Consortium of Isolated Halophilic Microorganisms. Environ. Technol.
31 (11): 1229–1239. http://dx.doi.org/10.1080/09593331003646612.
Sharghi, E. A., Bonakdarpour, B., Roustazade, P. et al. 2013. The Biological Treatment of High Salinity Synthetic Oilfield Produced Water in a
Submerged Membrane Bioreactor Using a Halophilic Bacterial Consortium. J. Chemical Technol. Biotechnol. 88: 2016–2026. http://
dx.doi.org/10.1002/jctb.4061.
Tellez, G. T., Nirmalakhandan, N., and Gardea-Torresdey, J. L. 2005. Kinetic Evaluation of a Field-Scale Activated Sludge System for Removing Petroleum Hydrocarbons From Oilfield-Produced Water. Environ.
Progress 24 (1): 96–104. http://dx.doi.org/10.1002/ep.10042.
Thunderhouse Instruments. http://thunderhouse4-yuri.blogspot.com/2012/
05/geotrichum-candidum (accessed 16 February 2014).
Arnold Janson is the principal engineer at the ConocoPhillips
Global Water Sustainability Center in Doha, Qatar. He has
been with the company for 3 years. Previously, Janson worked
28 years for Zenon Environmental and GE Water. His research
focuses on the application of advanced technologies for produced-water treatment and reuse. Jnason holds 20 US patents
plus a variety of international patents on the subject of membrane-system design and membrane bioreactors. He holds an
MS degree in chemical engineering from the University of
Waterloo and is a licensed professional engineer in the Province of Ontario.
Ana Santos is an engineer at ConocoPhillips Global Water Sustainability Center in Doha, Qatar. She has been in the company for approximately 2 years. Santos’ research interests are
in produced-water management for reuse and membrane systems. She has authored or coauthored more than 10 papers.
Santos holds a PhD degree in membrane-bioreactor technology from Cranfield University, UK, and an MS degree in biotechnology engineering from University of Algarve, Portugal.
Altaf Hussain is an engineer in ConocoPhillips Global Water
Sustainability Center. He has been with the company for 5
years. Hussain’s research interests include produced-water
management and thermal and membrane desalination processes. He has authored or coauthored more than 10 technical
papers and holds one US patent. Hussain holds a PhD degree
in chemical engineering from King Saud University.
Simon Judd has served as the Maersk Oil Professorial Chair in
Environmental Engineering at Qatar University since 2012 and
has been a professor of membrane technology at Cranfield
University since August 1992. He has overseen a number of significant membrane filtration and membrane bioreactor
research projects during the past 20 years, and within the past
2 years has embarked on research into produced-water-treatment technology. Judd has authored or coauthored 6 books
and approximately 150 publications in peer-reviewed journals.
He holds a PhD degree from Cranfield University, UK.
Ana Soares is a lecturer in biological engineering at the Cranfield Water Sciences Institute. She has 9 years’ post-doctoral
experience in academic and industrial research and development. Soares has published 27 peer-reviewed-journal publications and has published in 19 conference proceedings. Her
main areas of research include biological wastewater treatment and nutrients, recovery of nutrients as resources, hazardous pollutants, and sensor development. Soares holds a PhD
degree in environmental engineering from Lund University,
Sweden.
Samer Adham is the Manager of Water Solutions at ConocoPhillips. Before this position, he was the managing director of
the Global Water Sustainability Center in Qatar for 7 years.
Adham was also a vice president and the manager of MWH’s
Applied Research Department in Pasadena, California. His
research interests lie in the areas of advanced membrane
technologies for treating byproduct water from oil-and-gas
operations and for municipal and industrial water sustainability. Adham has authored more than 40 peer-reviewed
research papers on a variety of water-related topics including
reverse osmosis and nanofiltration and holds one US patent.
He holds a PhD degree in environmental engineering from the
University of Illinois, Urbana-Champaign.
2015 SPE Journal 7