| 
|
High-efficiency reverse osmosis, or HERO, not only enables ultra-purity on the inbound stream, but it can greatly reduce waste volumes, even down to zero-liquid discharge.
By David Engle
Phoenix, 2004: A major silicon chipmaker was struggling over how to make its extremely hard water ultra-pure, and this desert region needed lots of it: hundreds of gallons per minute (gpm). Anti-scalants and softeners were being added to help a conventional reverse-osmosis (RO) system mounted on leased trailer skids. But, despite the chemists’ determined efforts, total organic carbons (TOCs) at the back end were far too high, at 15 parts per billion (ppb) to 30 ppb.
Plant technicians called the vendor. “This isn’t working,” they said.
The vendor suggested giving up on conventional RO and trying instead a fairly new approach, high-efficiency reverse osmosis, or HERO. Almost immediately after the HERO installation, recalls the plant’s operator (who prefers that he and his facility not be named), TOC dropped from between 15 ppb and 30 ppb to below 8 ppb; eventually, TOCs were rinsing down even lower, to a consistent 4 ppb to 5 ppb in the HERO permeate. In the bargain, by abandoning conventional RO and building a HERO facility, the plant was able to get a sewer discharge permit to meet expanding manufacturing capacity requirements. HERO not only enabled a much higher-quality makeup stream, but it greatly reduced the system’s waste volumes. The first-pass HERO reject water—as intensely silicate-laden as it is—was reusable as scrubber makeup or for acid waste neutralization.
Ultimate result: In this difficult case in Phoenix—using the same feedwater volume throughout this time—the ultra-purified output post-HERO rose from 335 gpm in 2004 to 500 gpm in 2006: today it’s even higher. The plant manager reports there’s actually more water available for reuse than is needed for the manufacturing process. Plant engineers are planning to run a line to a cooling tower and acid scrubber to pipe the precious recovered water there, at which point reuse efficiency will soar even higher.
Hard-Water Problems, HERO “Solutions”
The name given to the process certainly describes what the system delivers: high-efficiency RO. HERO has been boosting water recovery and efficiency—often dramatically—ever since its inception as an industrial water treatment process (just prior to this decade) in an estimated 100 to 150 plants in 17 countries. About half—including such firms as Micron Technology, Intel, Seagate Technologies, ST Microelectronics, Samsung, and Sony, among many others—are making ultrapure water for microelectronics firms, like the one in Phoenix. Most of the others developed subsequently, in a second wave of completely different HERO applications designed for purifying concentrated municipal sewage and graywater in order for use in water-intensive steam power plants and miscellaneous other wastewater reclamation projects. Notable adopters include Bechtel, Calpine, and Komatsu.
Apart from the pure efficiency gains it achieves, the HERO process seems to be summoned into use most frequently to solve tough treatment challenges that refuse to yield to conventional RO. In the adoption also comes a typically rapid payback and a remarkably reduced need for RO membrane cleaning and replacement. Over time, the combined value of these benefits can amount to hundreds of thousands of dollars, while yielding impressively higher water efficiency. A few cases illustrate:
- Idaho: In the late 1990s, Boise, ID, became the very first HERO installation, for Micron Technology; the immediate benefit, as Micron Process Engineer Marty Lindgren recalls, was avoidance of a $5 million sewer discharge fee. That alone, he notes, would pay for several HERO plants. Over the eight years of operation since, RO membrane maintenance has been dramatically reduced. HERO is thus providing, says Lindgren, “the lowest cost of ownership of any system.” In 2007 Micron commissioned its second HERO, for a new plant in Lehi, UT, (read more below).
- The Philippines: A few years ago, an RO plant was struggling with debilitating silica concentrations of “up to 100 ppm in the feedwater, and getting only about 50% to 60% recovery—and that was really pushing it,” recalls Tom Chiara, an engineer with GE Infrastructure, based in Arizona. After commissioning its new HERO facility, the overseas plant is attaining recoveries of 90%-plus.
- Albuquerque, NM: In 2004 the US government installed its first HERO here, at the Sandia National Laboratory. After commissioning, performance and cost benefit were closely evaluated. Result: HERO indeed showed a higher efficiency, compared with conventional RO treatment, manifested in lower operating cost, reduced water and energy usage, and curtailed waste discharge volumes. Output water is purer, and the operating flux (gallons per square foot of membrane per day) is nearly double, according to a report from the Federal Energy Management Program (FEMP). Although HERO requires significant quantities of additive chemicals to achieve its gains, it does eliminate the need for antiscalants and RO cleaning chemicals. Additionally, the consequences of a membrane failure are minimized because the heavily treated water is more robust. The system’s reject water—being extremely softened—is reusable for assorted appropriate purposes, such as (in Sandia’s case) makeup water for an acidic gas scrubber and in acid waste neutralization.
- New Mexico: In 1999 a wafer plant owned by a major semiconductor manufacturer was facing a true water crisis, both in the sense of acquiring enough of it for a very water-intensive process and in treating what it had. Water demands of this chip-making giant made it the region’s biggest water user, extracting some 3,000 gpm. This stream carried an unusually high 65 ppm to 70 ppm of silica—compared to normal water with silicates at 5 ppm to 20 ppm, recalls a chemical engineer and water treatment consultant based in Silicon Valley, Debasish Mukhopadhyay, whom the plant owner had called for advice. Mukhopadhyay was told that several thousand RO plant membranes were also fouling so badly they needed cleaning every four to six weeks. Just as undesirably, conventional RO wasted 50% of the hard-water stream (compared to typical reject rate averages of around 25%).
“They couldn’t live with such costly waste,” says Mukhopadhyay.
Neither could Albuquerque’s business community, which regarded this huge volume of water usage—half of it draining as sewer waste—as a potential development killer. The plant owner tried various scale inhibitors and softeners, eventually succeeding in raising recovery to around 60%; that was better, but still not really good enough. The urgent necessity of getting much better results thus spurred a radical innovation by Mukhopadhyay: increasing silica solubility far beyond what had previously been achieved, by raising feedwater pH to unconventional heights. After months of development and testing, a prototype succeeded in boosting the water system recovery rate to an unprecedented 85%. And the onerous, labor-intensive RO membrane cleaning intervals, which had bogged down production month after month, were instantly extended to an unheard-of six to eight months. In 1999 Mukhopadhyay patented his new process and dubbed it HERO (a registered trademark that he owns).
High pH Makes Smooth Water
In conventional RO, pre-treatment with chemicals and filtration are two givens, as hardness is almost always present in some degree. Ignored and untreated, hardness—caused by calcium carbonates and magnesium—along with silica, silt, and bacteria, will foul and scale RO membranes surprisingly quickly. Conversely, removal of hardness and silicates enables membranes to work far more efficiently—i.e., with higher recovery rates.
 |
Photo: GE Water Infrastructure |
| Tanks run on a continuous cleaning mode to keep water recovery near 100%. |
Mukhopadhyay’s HERO water-softening approach basically comes down to a kind of seesaw swing between acidity on one end, produced via weak acid cation softening, which facilitates the removal of carbon dioxide and very high alkalinity, of pH 10 or 11, at the other. That softening and pH adjustment makes many good things happen. By contrast, conventional RO also often raises pH, sometimes to as high as 8; but, going much higher than this, and achieving the extraordinary benefits thereby, was thought to be unattainable, due the various negative consequences.
In the most common embodiment of the HERO process that Mukhopadhyay devised, the feedwater undergoes a weak acid cation (WAC) softening before the chemicals are injected to raise pH. When hydrogen ions are exchanged with hardness ions, a balanced hardness-to-alkalinity ratio is achieved during the process, improving WAC efficiency. Hydrogen ions also reduce the pH, causing the water stream to be acidic and converting much of the alkalinity to carbonic acid and carbon dioxide. Adding more acid at this stage converts remaining alkalinity to carbon dioxide. Degasification (i.e., removal of carbon dioxide and other gases) follows. Now the pH can be jacked up very high to, say, 10.5; this will raise the silicates’ solubility and destroy any biological elements. Next, the feedwater enters the RO membranes in such a softened condition that little or no scaling or fouling occurs. This allows HERO’s remarkable water recovery at 95% to 99%. Again, that compares to conventional RO at about 75%.
In this description, note that three elements characterize HERO; if any two occur in any order or combination, then it’s defined as a HERO under Mukhopadhyay’s patent: removal of hardness, alkalinity reduction, and degasification, prior to an RO operating at a pH of 8.5 or higher (the level at which the extraordinary benefit of this super-softening begins to be realized). On this last score, a pH ceiling of 11 is also set, that figure being, notes Mukhopadhyay, “the highest allowable under RO membrane manufacturers’ warranties.”
Self-Cleaning, Self-Soaping
Water recovery approaching 100% is thus one key benefit; the other, again, is that cleaning chores shift from being a recurring nuisance to an easy task, on a frequency approaching zero. The five-year-old HERO at Sandia National Lab, for example, has not required a single cleaning since commissioning, notes the FEMP report.
GE Infrastructure’s Chiara explains the underlying chemical reasons why. Raising pH, as HERO does, “to 10.5 or 11, when you think about it … is essentially … [what you do] when you make soap!” he says. “By increasing the pH, you saponify the lipids”—i.e., fats typically present as a part of bacterial cells. “You are making lye soap. So,” he says, “in effect we’re running the RO with its own surfactant … 100% of the time.”
Under such constantly “soapy” conditions, “all silt is easily dispersed…. We’re operating in a continuous cleaning mode,” he says.
Too, he points out, several HERO plant managers have discovered how to tune the pH so as to optimize this “self-cleaning” to the point of being able to run their plants almost without need for a shutdown, at least for cleaning. Notes Chiara, “In two-stage systems, with highly concentrated second-stage reject,” systems can operate almost indefinitely without offline cleaning.
Chiara has been associated with several of GE Infrastructure’s HERO plants, including recently commissioned ones in Brazil, the Philippines, and Singapore. Moreover, if the feed water happens to be good-quality “city water” to begin with, this usually eliminates the need for the usual multimedia filters found in conventional RO designs. The traditional pre-RO cartridge filters are placed upstream of the gas transfer membrane modules. And, he adds, the threat of biological fouling—easily arising in conventional RO—“is completely eliminated with HERO,” thanks to the pH of the process environment. The high pH environment of the RO eliminates bacteria that might otherwise have survived the low pH pretreatment.
Also on the topic of membrane cleaning, Micron’s Lindgren tells a remarkable anecdote. The company had once purchased an RO-equipped plant in Texas. Aged membranes in it were so badly fouled that they’d turned green with moss. The company was going to discard them as a loss, when Mukhopadhyay intervened. “No, no, no, no, no!” he insisted. “Don’t do that! Throw them into the HERO system!” So, Micron packed-up several hundreds slimy membranes and shipped them to Boise. Lindgren subjected them to the high-pH “cure.” After several days, “They were restored to good condition and operated for two more years,” he says, “with very good results.”
Operationally, too, a HERO provides, he says, a kind of natural “self-healing recovery.” This comes in handy if an RO line ever slips out of its proper pH, as can of course conceivably occur. Typically, the result would be scaling and potentially costly cleaning or replacements. But the HERO’s unique pH environment and pretreatment counteracts any such system upsets. “The membranes will clean up by themselves,” says Lindgren.
Running the HERO has proven to be, he adds, “fairly easy.” The trickiest process turns out to be the WAC exchanger (where high pH is of course not present). Early on in Micron’s HERO plant operation eight years ago, scaling with calcium sulfate sometimes occurred “and was blinding the ion-exchange beds, preventing good regeneration,” he recalls. Switching from sulfuric acid for regeneration to hydrochloride, which eliminated the calcium sulfate solubility problem, solved these difficulties.
These and other experiences that Lindgren and Micron gained with their first HERO have yielded a number of improvements in the newly commissioned second plant in Utah. One example is an innovative two-stage WAC. “In the second pass, the sodium acts as a polisher to make sure we get to an extremely low level of hardness feeding the RO,” Lindgren explains. “This makes the system even more robust. We don’t have any concerns anymore of hard water or calcium carbonate scaling the RO membrane. By comparison,” he says, “in a single-pass WAC, if you had a bed fail, and you did get some calcium or some carbonate leakage to your HERO, you could potentially have a scaling issue.”
Similarly, the newest HERO plows back the feedwater within the RO itself to get even higher performance and quality.
Cost-Saving Gains
The reduction in membrane cleaning, combined with water recovery approaching 99%, translates into serious money and a quick payback. Example: Mukhopadhyay describes a 500-gpm HERO treatment operation installed in Austin, TX. Owners found that the upgrade yielded 140 more gpm—meaning 200,000 more gallons per day, and 60 million gallons per year. With city water rates at $8.50 per million galllons, the annual savings on water alone come to over $500,000.
At Sandia, FEMP’s formal cost-comparative study of a hypothetical 250-gpm HERO and the conventional alternative—RO with electrodeionization (EDI)—showed that HERO would come out operationally costing about 25% less, head-to-head. Here’s the breakdown:
- Estimated utility costs with HERO, $251,000 [1,849,760 kilowatt-hours (kWh)], versus $404,000 (2,838,667 kWh) for RO/EDI
- Feedwater for equivalent output, 117 million gallons with HERO versus 165 million with RO/EDI;
- However, HERO requires considerably more chemistry—nearly 200 tons of acid and caustic treatment, costing about $57,716. This compares to 57 tons of antiscalants and other chemicals, for $37,269, using an RO/EDI system.
- Other consumables: $370,506 for HERO, $454,353 for RO/EDI.
- Net total difference, annually: $680,139 for HERO, $895,735 for RO/EDI.
- And note that these figures do not reflect the quite significant reduction in cleaning and maintenance with HERO.
At Micron’s Boise, ID, site, for instance, notes Lindgren, before the HERO was implemented, membranes needed cleaning every four to six months; HERO came along and stretched this to “18 to 24 months or longer—and we’re now doing 2,200 gpm, (i.e., producing much more water with the same membranes). The plant uses about 1,500 membranes (cost: about $500 apiece) totaling over a million square feet. The process of cleaning and rinsing takes batches of several dozen at a time, with a cleaning cycle lasting four hours. HERO has saved the company hundreds of hours of labor and, cumulatively, probably millions of dollars in extended membrane service life and reduced workload, he says.
Capital costs for HERO systems are something of a curiosity too. In sizes of 25 gpm to 50 gpm, these come out generally higher than conventional RO systems; however, beyond the 50 gpm, comparative figures shift in favor of HERO: Its higher flux allows installation of fewer RO membranes. “Above 50-gpm system size,” says Mukhopadhyay, “HERO is almost always cost-effective.”
The numbers will come out various ways, though, depending on whether the system is a retrofit or new plant from scratch. If the latter, it’s almost always a “slam dunk, economically,” says Lindgren, who has participated in costing out and running both kinds. The modestly higher first cost for a HERO will usually pay back within a few years, he suggests.
Chiara notes, too, that thanks to HERO’s higher flux and recovery, this upgrade can help a growing plant avoid the expense of a big RO expansion. Example: A plant in Singapore “needed to expand, and instead of building a new system, we just retrofitted the existing one,” he says. Water output was increased by 50%, without trucking in much new equipment or building more floor space. This strategy also saved Singapore’s water supply about 178,000 cubic meters annually. The system’s rejection of boron, which had been only in the 50% to 60% range, improved up to 98.5% to 99.5%. Silica passage was reduced ten-fold, to only 0.003% to 0.005%. (Note: These figures were reported at an ultrapure water conference, in a paper by William McClain, then vice-president of system supplier Ionics.)
In such economical capacity-improvement scenarios, adds Chiara, HERO is becoming a preferred solution worldwide.
 |
Photo: GE Water Infrastructure |
| High-efficiency reverse osmosis reduces the power generation industryís water needs. |
HERO for Zero-Liquid Discharge
Besides doing ultra-pure pretreatment, a second major market for this super-softening methodology has proven to be wastewater reclamation, particularly for the power generation industry. Electricity-making plants often require several hundred gallons per minute of water for cooling or steam. However, they’re often sited in the boondocks where water is in short supply, is problematic in quality, or is subject to environmental discharge restrictions.
As a preferred solution, sewage graywater, even with its often-high total dissolved solids, can be readily treated with HERO. Output can meet the power plant’s very exacting standards and volume needs, where any other treatment would likely be inadequate. A major water industry supplier, Aquatech, near Pittsburgh, PA, is the HERO licensee handling power industry application. Aquatech has several dozen such plants to date, typically in arid sites here and abroad.
A related and often similarly critical HERO benefit is its ability to recycle concentrated wastewater from, say, cooling towers or scrubber blowdown, to the point of achieving near zero-liquid discharge. This can prove decisive in allowing otherwise impossible-to-build projects to be approved, in places where sewer capacity is either non-existent or prohibitively costly.
Plants of this kind have been built worldwide, notes Mukhopadhyay. A typical example: At the Griffith Energy Project in Arizona in 2006, cooling tower blowdown volume at about 230 gallons per minute was coming out saturated with silica and assorted extreme hardness elements. Conventional membranes were not usable. Engineers turned to HERO to reduce the wastewater discharge volume by 90%—i.e., near zero-liquid discharge. Developers thus avoided paying for a thermal brine concentrator at more than twice the cost of the HERO; they also enjoyed “enormous” savings on the power needed to operate it, according to co-developer Aquatech.
Still another innovative case: A power plant proposed for arid central Mexico was initially denied permission and was refused access to the insufficient supply of local groundwater. Developers ultimately won a permit, though, by first building a sewage treatment plant for nearby San Luis de La Paz; its graywater from processed raw sewage could be piped to the power station for reuse. Here, microfiltration and conventional RO were both rejected as inadequate to the challenge. Instead, HERO now accomplishes demineralization, silica, and TOC removal and other conditioning. During this new Bajio Power Plant’s first two years of operation, membrane cleaning has been needed but once.
In other locales as well, permit standards have become so tough that a zero-liquid discharge design is a useful strategy simply to win quick construction approval. Cutting the red tape this way can cost-justify the application all by itself. With rigorous standards to overcome, notes Mukhopadhyay, “By the time you’ve made the water so good that you can discharge it, you might as well close the loop and make it good enough to reuse it again, instead of throwing it out.” The multiple reuses just become “gravy.”
Finally, after several years of working in a “high pH” frame of mind, Mukhopadhyay has more recently invented a mirror-image complement to HERO, which he calls XRO. Operating at low pH, this acidic “alterego” reclaims acidic wastewater, such as is found in acid mine drainage, metallurgical facilities, and semiconductor manufacturing. Rather than receiving pre-treatment, the XRO feedwater hits the RO membranes first and then undergoes extensive post-conditioning.
Mukhopadhyay sees as a certain irony in all this: acidic wastewater treatment; zero-liquid discharge applications; feeding thirsty power plants with reclaimed sewage graywater, for gosh sakes! “It’s all,” he says, “just diametrically opposite” to what he’d first envisioned.
But it’s said with a chuckle. His high-pH HERO has come a long way from those immaculate semiconductor plants, and he has no complaints.
David Engle, a writer based in La Mesa, CA, specializes in construction-related topics.
OW- November/December 2007
|
