Please fill in the name of the event

2022 SPE Western Regional

you are preparing this manuscript for.

Please fill in your 6-digit SPE

SPE-209256-MS
manuscript number.

Please fill in your manuscript title. Improved Reverse Osmosis Membranes for Treating Produced Water

Please fill in your author name(s) and company affiliation.

Given Name Middle Name Surname Company
Rich Franks Nitto - Hydranautics
Xiaofei Huang Nitto - Hydranautics
Craig Bartels Nitto - Hydranautics

This template is provided to give authors a basic shell for preparing your manuscript for submittal to an SPE meeting or
event. Styles have been included (Head1, Head2, Para, FigCaption, etc.) to give you an idea of how your finalized paper will
look before it is published by SPE. All manuscripts submitted to SPE will be extracted from this template and tagged into an
XML format; SPE’s standardized styles and fonts will be used when laying out the final manuscript. Links will be added to
your manuscript for references, tables, and equations. Figures and tables should be placed directly after the first paragraph
they are mentioned in. The technical content of your paper WILL NOT be changed. Please start your manuscript below.

Abstract
For many years, reverse osmosis (RO) elements have been used in the treatment of produced water,
including at several sites in California. The RO reduces salts and organics in the produced water to a level
that allows for disposal or reuse. The RO elements used to treat produced water are similar in chemistry
and construction to the conventional seawater RO membrane. But compared to seawater, the
characteristics of produce water are unique and varied. The conventional seawater membrane comes with
pressure and temperature limitations that restrict its ability to treat a wide range of produced waters.
Specifically, conventional membranes have a temperature limit and a pressure limit. Only a portion of
the produce waters needing treatment fall within the membrane’s temperature and pressure limitations.
Many produced waters, including produce waters associated with SAGD, require membranes that can
accommodate higher temperatures up to 60 C. Other produced waters may allow for treatment at ambient
temperatures but their higher salinities above 60,000 mg/l TDS require RO membrane to overcome high
osmotic pressures and operate at feed pressures up to 1800 psi.

In recent years, membrane manufacturers have enhanced their exiting RO elements to address the
challenges associated with the treatment of unique industrial streams such as produced water. Specifically,
new, more robust element construction allow designers to push beyond the normal limits of temperature
and pressure. One such element allows for operation at temperatures up to 90 C while a second, ultra
high-pressure RO (UHPRO), can concentrate the total dissolved salts (TDS) up to 120,000 ppm (12%)
while operating at pressures up to 1,800 psi (124 bar).

These unique elements can be used to increase the overall efficiency of the treatment facility by reducing
the cost of brine disposal and maximizing water recovery. This paper will show how these new elements
perform when operated beyond conventional pressure and temperature limits - including how individual
ion passage and water permeability are affected at extreme conditions. This paper will share element
performance data from laboratory and pilot studies. The data will be used as a basis for new designs at
the extreme conditions associated with produced water treatment.
2

Introduction
Produced water is that water which is brought to the surface as part of the oil and gas extraction process.
The produced water contains a high level of dissolved salts and other organic and inorganic
contaminants. Produced water composition varies as a function of geological formation, injection water,
and the hydrocarbon properties in various reservoirs. For steam flooding in a sandstone reservoir, as
commonly seen in California, produced water possesses unique challenges for treatment such as high
silica, high hardness, and high boron. Beneficial reuse involves stringent local regulations on essential
toxicity and key water parameters impacting the environment such as TDS, ammonia, and boron. The
California Regional Water Quality Control Board regulates the surface water discharge limit on TDS to
be less than 500 mg/L and on boron to be less than 0.5 mg/L. Not only does the poor quality and strict
regulations make produced water a challenge to dispose of or treat, but the large quantities make
handling difficult as well. The volume of produced water can be as much as ten times the volume of the
oil extracted.

Historically, the primary method for dealing with produced water is reinjection. A portion of the
produced water is injected into the oil producing zones to improve oil recovery through water or steam
flooding. Another portion of the produced water is disposed of through deep well injection. Deep well
injection is limited by the capacity of the injection wells which in turn limits the oil field productivity.
As far back as the 1990s, processes for reclaiming and reusing the produced water were researched and
piloted (Bartels, 1990). These processes were shown to be effective but were not initially applied due to
their high-cost relative to the low cost of oil. However, as oil prices and oil scarcity began to increase
during the first decade of the 21st Century, there was a renewed interest in processes to reclaim and reuse
the produced water and increase oil production. During the same period, technologies were also
improved to make treating produced water more economically attractive. Several treatment
technologies, including membranes, are used to treat produced water for environmental and agricultural
reuse. Among these technologies is desalination by reverse osmosis. Specifically, reverse osmosis
membranes are used as one of the final treatment steps after oil, grease, solids and hardness removal and
pH elevation. The RO step is designed to remove the remaining dissolved salts and organics, including
sodium, silica and boron.

Produced water characteristics

The specific characteristics of produced water vary depending on the age and geology of the formation,
but most produced waters contain contaminants that pose a challenge to the successful long-term
operation of the RO. These contaminants include suspended silt and clay, suspended oil droplets,
dissolved organics such as acetic acid and acetone, dissolved gases such as hydrogen sulfide, heavy
metals, and sparingly soluble salts. The temperature of the produced water can also pose a challenge to
the RO system. The practice of steam injection can increase water temperatures to as high as 80 C while
the standard RO membrane is limited to 45 C. Table 1 below compares the produced water
characteristics from two different sites in North America (Nagghappan,2006).
 3

Table 1. Example of produced waters.

Site 1 Site 2
(New Mexico) (California)
pH 7.0-7.5 7.5
Temperature (C) 80 85
Calcium 620 80
Magnesium 110 10
Sodium 5,088 2,300
Potassium 95 39.1
Strontium 15 1.0
Bicarbonate 561 670
Carbonate 0.6 -
Sulfate 2,150 133
Chloride 7,800 3,400
Silica 100 240
Boron 3.5 26
Total Dissolved Solids (TDS) 16,000 7,000
Total Suspended Solids (TSS) 55.0 10
Soluble Organics 30 80
Methyl Ethyl Ketone 1.0 -
Acetone 5.0 -
Oil and Grease 10-50 20 ~ 30

RO System Design

Extensive pretreatment of the produced water is required for stable operation of the RO. A number of
water treatment technologies are used to reduce oil and grease, hardness, metals, temperature, and
suspended solid before going to the RO. At an oil production facility in Central California, pretreatment
before the RO is used to address the following (Franks, 2009):

• Free Oil. Free Oil concentrations as high as 100 ppm are reduced using induced gas floatation
(IGF) , and walnut shell filters to levels less than 1.0 ppm.

• Hardness and heavy metals. Softening and metal removal is achieved through chemical softening
and settling and ion exchange softening. Hardness levels are reduced below 0.1 ppm as CaCO3.

• Temperature. It is necessary to reduce temperatures using a heat exchanger. Temperatures as high

as 80 C are reduced below 40 C before sending to the RO system using conventional elements.
However, new element construction discussed in this paper would reduce the need for temperature
reduction.

• Suspended solids. Suspended solid concentration in the raw water to the RO system can exceed 50
ppm. With multi-media filters, the suspended solids concentration is reduced below 0.5 ppm. The
Silt Density Index, a measurement of colloidal and particulate fouling potential, is less than 4.0 for
stable RO performance.
4

Despite extensive pretreatment, the presence of dissolved organics means the average RO system flux
treating produced water is similar to a RO treating municipal wastewater at 10 gfd to 12 gfd. The
conservative flux reduces fouling and cleaning frequencies but does nothing to achieve higher
recoveries.

System recoveries can range from 65% to 90% depending on the TDS and the scaling potential of the
feedwater. This limitation can be problematic when reinjection capacity of the brine stream is limited
by the injection well capacity. The TDS (and associated osmotic pressure) increases with increasing
recovery. In the final element, the net driving pressure (NDP) should be sufficient to overcome the
osmotic pressure of the concentrated water and produced a permeate flow. Going to a higher recovery
results in the absence of permeate flow from the last element unless a higher feed pressure can be
applied to the front of the RO system. But the max feed pressure is limited by the construction of the
RO element to less than 1200 psi. Newly designed elements discussed below, allow for the use of
greater feed pressures up to 1800 psi.

The ability to design for higher recoveries is also limited by the scaling potential of the feedwater. The
removal of hardness by one of the pretreatment steps, reduces the potential for CaCO3 scaling. The
hardness removal also allows for an increase in pH, which reduces the potential for SiO2 scaling.

Enhanced RO membranes for higher salinity and higher recovery.

As stated above, the pressure limitation on a standard RO element effects the amount of permeate that can
be recovered from an RO system treating produced water by limiting the amount of osmotic pressure that
can be overcome in the last element. Standard RO elements are limited to a maximum feed pressure of
1200 psi. The maximum pressure limitation is temperature dependent so that the max pressure decreases
from 1200 psi as temperature increases above 25 C. (Figure 1). Depending on the materials of
construction, this pressure vs temperature limitation is governed by either the potential for sudden
mechanical failure or a gradual flux loss due to compaction of the membrane and its support layers.
Regardless, some membrane manufacturers now offer ultra-high pressure RO (UHPRO) elements which
overcome the restrictions of standard elements and allow feed pressure up to 1800 psi. The new UHPRO
elements use more robust materials to prevent mechanical failure and reduce flux loss associated with
compaction. As with standard elements, the maximum pressure limitation decreases with increasing
temperature to mitigate the effects of compaction at higher temperatures (Figure XX). The maximum
pressure limit of the UHPRO remains higher than conventional RO up to the maximum temperature of 45
C.
 5

Temperature - oF
32 52 72 92 112 132

140 2000
Maximum Pressure - bar

UHPRO

Maximum Pressure - psi

120 1800
1600
100 1400
Conventional RO
80 1200
1000
60 800
40 600
400
20 200
0 0
0 5 10 15 20
25 30 35 40 45 50
Temperature - oC
Figure 1. Maximum Pressure vs Temperature limit for UHPRO and Conventional RO

Most of the flux loss in a conventional RO system is caused by fouling with only a very small contribution
from compaction. A majority of the flux loss caused by fouling is reversible through cleaning. However,
when operating UHPRO at the extreme pressure limits, compaction contributes more to flux loss. The
portion of the flux loss caused by compaction is irreversible. This flux loss is a function of both the
maximum operating pressure and maximum operating temperature. In other words, the higher the pressure
and the higher the temperature, the greater the compaction. The rate of flux loss is greatest during initial
compaction period which occurs instantaneously and then stabilizes within the first 24 to 48 hours of
operation.

In a controlled laboratory setting, with a pure, non-fouling, sodium chloride feed of 8.5%, the membrane
produces an average flux of 10 GFD when pressurized to 1740 psi. Apparent flux loss over a 48-hour
period is minimal. However, when the element retested at standard test pressures of 800 psi, the element’s
permeability is reduced by 50% relative to before the 1740 psi exposure test. When operating in a full
scales system, such a flux loss impacts the flux distribution between the lead element and tail element
within a pressure vessel and ultimately determines the maximum achievable brine concentration and,
therefore, the maximum achievable recovery. Depending on the ionic composition of the feedwater, the
brine concentration may reach beyond 120,000 mg/l before the osmotic pressure approaches the pressure
limit of 1800 psi and the resulting net driving pressure is insufficient to produce any flux out of the tail
element.

Further testing done in a laboratory setting demonstrates ability of the UHPRO to concentrate the salts
beyond that of the conventional RO element (Gisclair, 2021). Using a pure sodium chloride feed to
eliminate the potential for any fouling, a single element was run at 1740 psi and 25 C while the permeate
was sent to drain and the concentrate was recycled. The results shown in Figure 2 replicate the increasing
feed salinity along the length of a pressure vessel from the lead element to the tail element. At the
beginning of the test, when replicating the lead position, the element produced the highest flux of 25 LMH
(14.7 gfd) and the best permeate quality of less than 500 us/cm. By the end of the test, as the reject stream
approached 126,000 mg/l of sodium chloride to simulate the osmotic pressure at the tail position, the
permeate flux decreased to 2 LMH (1.2 gfd) while the permeate quality increased to 5500 us/cm.
Lead Position Tail Position
6

Figure 2. Single UHPRO laboratory test at 1740 psi (120 bar) and 25 C when concentrating a 65,000 mg/l
NaCl solution. (Gisclair, 2021).

The controlled laboratory testing characterizes the UHPRO element performance in the absence of any
fouling. But based on RO systems currently treating high fouling produced water, it is assumed that a
combination of fouling and compaction will lead to greater flux loss in the UHPRO. The flux loss
associated with both fouling and compaction was carefully monitored in a pilot study using UHPRO to
treat cooling water blowdown (Mandel, 2021). The raw water was softened, treated by ultrafiltration and
then concentrated up to 82,429 mg/L TDS by standard seawater RO membranes. This seawater RO
concentrate was then sent to the UHPRO and further concentrated up to 119,167 mg/l. Figure 3 shows
the membrane UHPRO flux during the first six days of pilot operation. Flux varies hourly as the pilot
operates in recycle mode and the original feed salinity is concentrated up. After an hour in recycle mode,
the concentrated feed is discharge and new feed is reintroduced. In the first three hours of operation at
1494 psi (103 bar) and 25 C, the membrane saw its greatest flux loss, decreasing 45% from a max flux of
32 gfd (55 LMH) to 18 gfd (30 LMH). Within the first 24 hours, the membrane stabilized at a max flux
of 10.6 gfd (18 LMH) - a 67% reduction. After flux stabilization, the temperature was increased to 30 C
and the feed pressure was increased to 1653 psi (114 bar). No further flux decline occurred during the
remaining five days of pilot operation.

After completing the pilot study, the element was evaluated and found to have good mechanical integrity.
Retesting of the element at its original standard test pressure of 800 psi confirmed the 67% flux loss seen
during pilot operation. The element retest also showed rejection improved from 99.80% at original factory
test to 99.86%. The improved rejection was caused by organic fouling.
 7

Figure 3. Single UHPRO laboratory test at 1740 psi (120 bar) and 25 C when concentrating a 65,000 mg/l
NaCl solution. (Mandel, 2021)

The combination of laboratory testing and pilot testing with the UHPRO sufficiently characterizes the
performance of this membrane and allows designers of an UHPRO treating produced water to project the
required feed pressure after compaction and fouling occurs. If precipitation of organics and salts can be
controlled, the use of UHPRO for treating produced waters allows the already concentrated TDS from the
conventional RO to be concentrated further to almost 130,000 mg/l and allows an increase in the overall
recovery of the RO system.

The obvious downside to operating with UHPRO is the increased energy consumption. But the high
energy can be partially offset by applying existing energy recovery technology in new ways. Specifically,
the use of multiple energy recovery devices in a multistage RO means the ultra-high brine energy from
the last stage can be proportionally and efficiently distributed between more than one stage. This concept,
known as BiTurboTM, is shown in Figure 4. A version of the BiTurboTM design has been proven on two-
stage seawater RO systems and can be applied to RO systems treating produced water (Gisclair, 2021).
For a multi-stage RO treating produced water, the initial stage(s) would use conventional RO elements.
Depending on when the pressure limit of the conventional RO is reached, the later stage(s) would use
UHPRO. Two turbos would be installed to boost two separate stages. Most of the energy would be used
by the first turbo (Turbo 1) to boost the final stage. After the Turbo 1, there would still be sufficient
energy remaining for Turbo 2 to boost the first stage and therefore reduce the energy consumption of the
high-pressure pump (HPP) on the feed stream. To illustrate the multistage turbo design more clearly,
Table 2 lists the projected flows, pressures, and salinities associated with the design in Figure 4. The
projected results are based on treating a high salinity feed of 46,371 mg/l TDS at 25 and achieving a
recovery of 60%. In this design, Turbo 1 transfers the energy from the UHPRO brine (Stream 6) to the
feed of the UHPRO (Stream 4) resulting in a boost pressure of 500 psi. The exhaust pressure from Turbo
1 is still at 575 psi and is used in the second Turbo 2 to provide a boost of 160 psi to the feed of the first
stage. The use of the turbos provides good flux balance between stages with an average flux in the first
stage of 9 gfd and an average flux in the second stage of 6.3 gfd.
8

Conventional RO

HPP

Turbo 2

UHPRO

Turbo 1

Figure 4. Multistage turbo design with conventional RO membranes in the first stage and UHPRO
membranes in the second stage.

Stream No. Flow (gpm) Pressure (psi) TDS (mg/l)

1 1540 0 46372
2 1540 829 46372
3 1540 989 46372
4 838 979 84967
5 838 1479 84967
6 616 1463 115115
7 616 575 115115
8 616 33.9 115115
9 701 0 271
10 222 0 1365
11 924 0 534

Table 2. Projected flow, pressure, and TDS of each stream in the multistage turbo design shown in
Figure4.

Enhanced RO membranes for higher temperatures

Produced water coming out of mature steam floods has temperature as high as 90 °C, which requires
significant cooling if commercial RO treatment processes are considered. The high silica and petroleum
organics in such produced water, on the other hand, would prefer a warmer temperature procedure
which reduces the risks on fouling to the heat changer and RO membrane. Conventional RO membranes
are limited to a maximum temperature of 45 C. Therefore, high temperature RO membrane attracts
special interest in treating produced water.

Boron and silica rejections with high temperature membrane were studied in the lab using synthetic
produced water and real produced water collected from the field. Hydranautics PRO-XT2 membrane
was used in this study. Salt rejection is above 99.7% when treating a standard test solution of 32,000
mg/L NaCl at 800 psi applied pressure. All the laboratory tests used a single element system layout.
 9

Operating conditions were controlled at fixed design parameter settings. Laboratory synthetic water
composition is listed in the Table 3 below.

Analyte Concentration, mg/L

Calcium 10.1-12.7
Magnesium 0.17-2.6
Sodium 3120-3310
Potassium 103-126
Chloride 4300-4533
Sulfate 192-205
Alkalinity as CaCO3 622-1134
Boron as B 77.7-85.7
pH 10.9-11.3
TDS (Total dissolved solids) 8450 -9219
Table 3: Laboratory synthetic water used to characterize high temperature RO elements.

Boron and TDS rejections were evaluated with synthetic water tests using a single element at
temperatures between 25 °C and 60 °C. Test results (Figure 5) were compared with projection software
simulation results which revealed the accuracy of TDS rejection prediction by the software. The boron
rejection was underestimated by the software (Figure 6).

Figure 5 TDS rejection vs temperature for synthetic water, permeate flux was at 10.5 GFD, recovery
rate was at 15%.
10

Fig 6. Permeate boron concentration vs temperature for synthetic water, permeate flux was at 10.5 GFD,
recovery rate was at 15% at raised pH 11.

Tests with real field produced water (Figure 7 & 8) produced similar boron rejection results as from
synthetic water. Sodium rejection was observed to be significantly lower than chloride rejection, which
is attributed to the “swelling” effect at higher pH that increases hydroxyl ion passage and “pulls” sodium
into the permeate. [Franks et al, 2009]. Field produced water test gave lower sodium rejection than
synthetic water, which is believed to be due to the influence of petroleum organic compounds in the real
field water that can exacerbate the membrane swelling effects [Chen et al, 2021].

Figure 7. Boron, SiO2, Na and TDS rejection at different pH in the field produced water tests at 55 °C;
permeate flux was at 15 GFD, and system recovery rate was at 10%.
 11

Figure 8. Rejection comparison between synthetic water and field produced water at pH 11, 55 °C.

Membrane projection software was calibrated using synthetic water test results on boron. Full scale
system design was carried out using calibrated software. Three recovery scenarios were examined at
40%, 60%, and 75% using lab skid with close-loop batch filtration mode with single element (Figure 9
& 10). This assumed a treatment feed flow capacity at 44 m3/h using a 4inch PRO-XT2 element. Design
pH was at 11 and water temperature was set at 55 °C. Lab permeate boron measured at 75% recovery
operation condition was 1.24 mg/L. This low boron level allows an easy post treatment options such as a
secondary pass RO or polishing IX resin to satisfy the final boron target at 0.5 mg/L.

Figure 9. Boron, sodium, and TDS removal rate for high recovery single element tests with field
produced water at pH 11, 55 °C.
12

Figure 10. Full scale system boron and sodium rejection prediction vs actual lab results using single
element simulation.

Full scale system projection results on boron and sodium rejections were compared with lab single
element test results. Good matching was found on lab and simulation results with lab rejections slightly
higher than the program. A visual comparison of the water samples collected from the test can be seen in
Figure 11.

Figure 11. Water samples collected from the feed, permeate, and reject streams at 75% recovery test
with field water.

Conclusion

The use of standard RO elements for the treatment of produced water was originally piloted more than 30
years ago. For more than 15 years, RO elements have been used in full scale systems in California to treat
produced water. But the pressure and temperature limits of the standard RO element have created
limitations on the salinity and temperature of produced water than can be treated. But newly designed RO
elements, with new, more robust materials of construction, have been developed to treat a wider range of
industrial feedwaters, including produced water. These elements have pushed the temperature limits from
45 C up to 60 C. They have also pushed the pressure limits from 1200 psi up to 1800 psi.

Using the new RO elements for the treatment of produced water requires a clear understanding of how
permeability and ion passage are affected when operating at the extreme pressures and temperatures. In
the case of permeability, designers should expect a flux loss of more than 50% due to compaction.
 13

Additional flux loss will be caused by the high fouling nature of produced water. Projected feed pressures
should account for flux loss from both compaction and fouling.

In the case of ion passage, ultra-high pressure has relatively little impact. But operating at higher
temperatures significantly effects ion passage, especially the passage of boron. Projected permeate
quality should account for the higher temperatures over a wide pH range. The lab tests and software
predictions demonstrated encouraging results on high temperature RO membrane treating produced
water. At the examined condition of 55 °C and pH 11, boron rejection was greater than 98% while
sodium rejection was lower at 95%.

Acknowledgement
Part of the UHPRO and high temperature RO testing were collaboratively done with Saltworks
Technologies Inc. and Chevron Technical Center USA. The authors are grateful to collaborators’ works
and contributions.

References

Bartels, C. and Dyke C., (1990), “Removal of Organics from Offshore Produced Water Using
Nanofiltration Membrane Technology”, Environmental Progress (Vol 9, No. 3) pp183-186.
https://onlinelibrary.wiley.com/doi/10.1002/ep.670090320

Chen, C., Huang, X., Prakash, P., Chilekar, S., Franks, R., (2021). “Produced Water Desalination Using
High Temperature Membranes”. Desalination. 513 (2021) 115144.

Franks, R., Bartels, C., Anit, A., Nagghappan, L., (2009). “RO Membrane Performance When
Reclaiming Produced Water from the Oil Extraction Process”. International Desalination Association
World Congress, Dubai, UAE, 2009.

Franks, R., Bartels, C., Nagghappan, L., (2009), “Performance of a Reverse Osmosis System when
Reclaiming High pH - High Temperature Wastewater”. International Water Conference, Orlando, FL, 4-
8 October. https://www.proceedings.com/07490.html

Gisclair, M. et al, (2021), “Established Practices, Innovative Designs, and New Products: A New
Generation of Ultra High Pressure Desalinization”. American Water Works Association – Membrane
Technology Conference. West Palm Beach, FL.19-22 July.

Mandel, D., (2021), “99% Recovery of Scaling Cooling Tower Blowdown with a Reverse Osmosis
Membrane Demonstration Plant”. International Water Conference, Scottsdale, AZ, 7-11 November.
https://eswp.com/water/conference-archives/iwc-proceedings-2021/

Nagghappan, LNSP. et al, (2006). “Desalination of Produced Water Using OPUS Technology”.
International Water Conference, Pittsburg, PA 22-26 October. https://www.proceedings.com/01724.html