Advances in Membrane Technology for Beverage Water Treatment
Activated Carbon: Beverage Water
Treatment: Filtration: Membrane Technology: Reverse Osmosis
INTRODUCTION
Reverse osmosis (RO), ultrafiltration
(UF) and nanofiltration (NF) are relatively new words
to the beverage industry, describing a series of filtration processes that have
recently found their way into quite a number of production facilities. While
all three tend to sound as if they describe new high-tech, very difficult to
understand processes, in fact, the concepts are really quite easy to
understand. All three simply describe forms of very fine filtration that are
capable of producing a final water product of a purity
unattainable through the more conventional means of water treatment.
FILTRATION CHARACTERISTICS
While the terms RO, NF and UF tend to set
each other apart, it is important to underscore just how physically similar
they are. All three are "membranes" created by coating a thin layer
of a very porous polymer, or plastic, onto a backing material. The end result
is the finest form of filtration presently known, with reverse osmosis being
the smallest, nanofiltration being a slight step
larger and ultrafiltration being a bit larger again.
The actual size of the pores in any of these polymeric type membranes is
measured in terms of Angstroms, with one Angstrom being one 10 billionth of a
meter, which is, of course, very small.
To put this in an easier to understand
perspective: consider if one square foot of membrane were the size of the
entire Pacific Ocean, a reverse osmosis pore would be roughly the size of a
dime. Nanofiltration would be about the size of a
half-dollar and ultrafiltration would be similar in
size to an old silver dollar.
Or perhaps more graphically: if you had the
opportunity to observe an RO membrane under a Scanning Electron Microscope
(SEM) at 8000 x magnification, the pores would be undetectable (Figure 1). This
is in contrast to the pores of a submicron filter
(0.2 µ), which are very distinguishable.
While each of the membrane types are very similar physically, the reasoning behind the
different labels is each of the three performs very different functions.
Reverse osmosis, for example, "filters" most inorganic salts from
water, allowing only the water to pass through the pores. This membrane will also
reject some forms of non-ionic organic compounds such as a fructose molecule
(molecular weight = 180). However, it will pass, quite readily, smaller
organics such as ethyl alcohol (mw = 46).
As the pore size is increased, larger and
larger organic compounds pass through, as does more of the salt. At a point
where all the salt passes, the membrane is then referred to as ultrafiltration. At that point only organics of greater
than 1000 mw are rejected.
Nanofiltration covers an area in-between with organic rejections of
300 -1000 mw and salt passages of 15 - 90%. Several different types of nanofiltration cover the entire spectrum of the area
between RO and UF.
Because UF has larger pores, it is the least
costly to operate, requiring fewer membrane elements and lower operating
pressure (and thus less horsepower) than the other two processes. As the pores get smaller, as they do for NF and RO (in that order),
the cost of operation increases somewhat. However, it is important to note
that the capital costs tend to be relatively similar since a large portion of
these costs is associated with piping, frames, controls, etc., all of which are
related to the quantity of water to be processed-- not the extent of
filtration.
Generally, UF is used when the only issues
are fine filtration and removal of most organic contaminants. RO is used when
the reduction in inorganic salts (e.g. sodium, alkalinity) is added to this
list; and nanofiltration is used when the interest is
only a partial reduction in the inorganic salt levels, along with fine
filtration and organic removal.
A good example of the different membrane
processes abilities can be created by simply filtering a sample of typical surface
water that is fairly heavily laden with organic contamination. In a trial a
sample of
While this does not offer assurance that low
molecular weight compounds such as trihalomethane
compounds (THMs) are affected, it does show that most
of the organic "precursors" are removed, thereby reducing the THM
formation potential with organic contamination being removed to a degree well
beyond more conventional means of water treatment. Such filtration has proved
to be a savior in cases where the raw water supply becomes unexpectedly
contaminated, such as occurred in
SYSTEM DESIGN
Because all three-membrane processes are
"dissolved" solids processes, the system must be designed in a manner
so that the contaminants are continuously removed before they become
"insoluble" solids. This sets the processes apart from other forms of
"normal" filtration where the intent is to remove contaminants either
in the form of a solid sludge or impinged within a disposable type of filter
media. Thus, an inherent aspect of a cross flow membrane system design is the
discharge of a small percentage of water carrying away the contaminants.
Naturally, as water is filtered, the
remaining contaminants become increasingly concentrated. The degree of
concentration is dependent on the ratio of water purified versus the water
discharged. If, for example, one sets this ratio at 50/50, the concentration
factor is 2x (meaning that all contaminants that do not go through the membrane
are concentrated by a factor of two). If the ratio is 75/25, the concentration
factor is 4x; at 90/10, 10x and so forth. The challenge is to set this
discharge as low as possible, while avoiding the concentration of any
constituent to a point where it precipitates, or changes from a dissolved solid
to an insoluble solid. Because each chemical species is different, the limits
of solubility for each compound must be studied separately.
In some cases, such as calcium carbonate, it
is possible to push the system beyond the limits of precipitation by adjusting
the pH to the more acidic range, converting the calcium carbonate to calcium
bicarbonate which has a higher level of solubility.
In a typical water supply, the limiting
factors are usually calcium sulfate, which precipitates at 2000 ppm; calcium carbonate, which will still precipitate at
levels in the 500-1000 ppm range even after pH
adjustment and silica precipitating in the neighborhood of 120 ppm.
Table 1:
|
Case study: |
|
|
Filter Diameter: |
72 inches |
|
Carbon Capacity: |
150 cuf
ft |
|
Design Flow: |
1 gpm/cu
ft |
|
Chloramine Feed: |
2.0 ppm |
|
Chloramine Discharge: |
0.0 ppm |
|
Activated Carbon Life: |
7 years |
Other factors such as barium (which has a
solubility limit of 2.0 ppm when in the presence of
sulfate) and even general dirt and debris found in the water must also be
considered since almost all contaminants are concentrated in the process. On
the other hand, in cases where all or a portion of the material does pass
through the membrane, as salts do with UF and non-ionized inorganic compounds
(such as silica) will do with NF, the concentration effect will be lower.
The upshot is that when one is processing
cleaner water supplies, such as water from the Great Lakes or from East Coast
surface supplies, it is not unusual to operate a membrane system with a low
percentage of water discharge (10% for example). But, when operating on water
supplies from the
A case in point is the comparison of the
system in
Table 2:
|
|
||||
|
|
City Water |
Permeate |
Discharge |
DuPage
|
|
Calcium |
104 |
12 |
572 |
288 |
|
Magnesium |
52 |
4 |
232 |
344 |
|
Sodium |
18 |
10 |
56 |
32 |
|
Alkalinity |
112 |
8 |
16 |
132 |
|
Sulfate |
37 |
7 |
754 |
411 |
|
Chloride |
25 |
11 |
90 |
121 |
|
TDS |
174 |
26 |
860 |
664 |
|
pH |
6.6 |
4.6 |
4.8 |
6.2 |
(All
figures. except pH, as CaCo3)
While this discharge is a
"waste water," it is important to keep in mind that it is
still good quality water-- the dissolved material is simply concentrated to a
higher degree.
Thus, it is very worthwhile to consider
using this water within the plant for container rinse, can warming and even in
some cases as cooling water. And, any water that does ultimately need to be
discharged from the plant can often be sent to the storm sewer system as has
been demonstrated by a
ACTIVATED CARBON
While membrane systems have found their way
into many production facilities, so far they have generally been limited to
replacing the more conventional means of alkalinity reduction and filtration
portions only; most plants continue to use activated carbon (AC) after the
membrane system as a means of removing chlorine. Thus, many of these
installations have combined a more conventional means (AC) with the newer
technologies (RO, UF and NF) which has created some new design considerations
for the activated carbon portion of the system.
Putting it simply: one must "throw away
the book" on AC when it comes to designing a system with activated carbon
operating downstream of a membrane system. Most of the old "rules of
thumb" relating to AC simply no longer apply. This is because so many of
the AC operating conditions are linked to the eventual fouling with dirt and
organic compounds-- all of which are absent in the water treated by a membrane
system.
Taking the fouling by organic compounds and
other foreign materials out of the equation sets up a condition where the AC
will last for a long time. Under these circumstances, the limiting life factor
likely becomes the simple breakdown of the AC itself to the point where the
pressure drop becomes intolerable. And, one must keep in mind,
the AC need not be backwashed nearly as often because the membrane filtered
water is cleaner. This minimizes the physical damage to the AC particles.
Tab. 3:
|
Case
Study: |
|
|
Filter Diameter: |
84 inches |
|
Carbon capacity: |
228 cu ft |
|
Design Flow: |
1 gpm/cu ft |
|
THM Feed: |
21 ppb |
|
THM Discharge: |
0.0 ppb |
|
Chlorine Feed: |
4.0 ppm |
|
Chlorine Discharge: |
0.0 ppm |
|
Backwash Frequency: |
once per month |
|
Activated carbon Life: |
13 months |
This is in contrast to the relatively high
particle loading from a conventional system which has been shown to send almost
260 million particles per liter to the activated carbon treatment step.
A case in point is a bottling plant in St.
Louis that has processed almost several million gallons of water over the past
seven years through an RO machine and then through the same charge of AC, never
witnessing a chlorine breakthrough. And, the feed to the system contains 2.0 ppm chloramines, not simply a small amount of chlorine. In
fact, had the city water contained simple chlorine, the loading to the AC would
have actually been less since the RO rejects about 30% of the chlorine,
yet passes 100% of the chloramines.
Part of the improved performance of AC on
chlorine compounds is related to the lower pH inherent to the membrane system
product water. Chlorine is more susceptible to reduction to chloride in more
acidic conditions. However, one must also be cautious to use a "low
ash" type carbon to avoid the dissolving of carbonaceous ash contaminants
when operating on acidic water.
The sizing for the St. Louis AC unit was a
very conventional one gpm per cubic foot of AC;
however, testing shows that AC can be very effective for very long periods of
time operating at flow rates as great as 3 gpm per
cubic foot.
Meanwhile, in
Had the same water supply been subjected to
the chlorination processes of a conventional system, the THM levels would have
a very strong potential of increasing significantly.
FUTURE
Long term, the impact membrane systems will
have on soft drink water treatment will very likely go beyond the simple
reduction of sodium levels or other inorganic contaminants. No doubt, the value
of using a membrane system will also include a more highly filtered water, THM
reduction, greater control for a more consistent water supply and the ability
to take on opportunities presented by "New Age" drinks that may be
the products of the future.