Introduction & Question: Antiscalant for small RO systems and question on RO membrane pressure 3

[gerhard_za]

Hi all,

I am brand new to this forum. I have been working in the water treatment industry for many years as a project manager. Of late, I have been forced into the process engineering side of the game and realise that there are a lot of critical aspects of water, especially reverse osmosis, that I desperately need to build expertise in. I am hoping you folk could lend a friendly hand!

The background is that there is a network of laboratories that require deionized water. I took on the challenge to supply them with a more robust system that is a bit more industrial than they're used to, but in favour of availability, with less intervention required.

I settled on using 4040 tap water RO membranes as they are by far the most common. 4021 and 2540 are less common and don't cost much less. I could not use a large multistage type pump as it is too bulky and noisy for the small lab environments. Thus I settled on 800GPD diaphragm pumps which give ±4LPM at ±5Bar. Feedwater TDS is <300PPM in 95% of cases so we do carbon and 5micron pre-treatment and push it through the RO directly. Then it runs through a mixed bed resin to remove the remaining ions. These plants are small and there are many (almost 200). So doing antiscalant with affordable dosing pumps require dilution and too frequent replacement thereof.

I have two questions (for now)
1) A typical 4040 membrane can give 2400GPD which is much less than my pump's 800GPD. I am only getting about 1200GPD from my setup with some back pressure but with lower conductivities they do fine. My initial reasoning was that more surface area allows for longer life but I also read that I actually need to apply more pressure to get better rejection. But higher pressure at low TDS input water means permeate flows in excess of 50% of feed water (and they claim on a single pass you should be looking at 17% to 30% recover).
*What am I missing in terms of planning for membrane life? DO I need to go smaller on the membrane so the pressure increases from the same pump? But increased pressure also introduces higher probability of scale formation?
2) As mentioned, the antiscalant or SMBS dosing is not an option. I don't have the space and have to settle for shorter lifetimes on RO membranes (regarding them as more of a consumable, due to extra hardware and consumable cost more or less surpassing the cost of a new one on an annual basis).
*Have any of you had experience with the Siliphos crystal balls or similar which I can throw into a cartridge and allow to dissolve on the way into the membrane? Is this suitable? Will it have more pros than cons? Their main ingredient is P204 or P205 (

https://carbonwaters.org/wp-content/uploads/2018/02/siliphosdatasheet.pdf)

I hope some of you have the time to read this and offer some advice. I am in a tight spot to make almost 200 systems in the field cheaper to maintain and prolonging membrane life in a small setup is key.

Kind regards
Gerhard from South Africa

 

[bimr]

As you are aware, there are trade-offs between operating at different flow and pressures. At the maximum flow and pressure, the membranes will not last as long for a variety of reasons. There will also be more water wasted when operating at low recoveries. Higher flow and pressures will result in more fouling and scaling as well.

With 300 different installations having different water quality and operators, you should expect to see widely variable performance.

I strongly recommend against the use of siliphos as these chemicals are known to cause scale in membranes. The benefits of phosphate will be marginal at best in most applications where phosphates are used for corrosion control but adverse results are to be expected with membranes.

 

[gerhard_za]

Thank you so much @bimr

At least knowing that Siliphos is not an option, allows for me to focus on other aspects.

Also, you are saying that lower pressure and flow in the membrane is not a bad thing. The question just beckons how high my pressure should be to get good performance vs. not ramping it up so high that I increase scaling potential.

 

[bimr]

Lower pressure and flow are a trade-off with recovery. Higher pressure will give higher recovery and require less membrane area. You are comparing the cost of the membrane to the cost of the wasted water (with lower recovery).

You may want to consider contracting with a service firm that provides water treatment services.

 

[ashtree]

You appear to have two issues:
1) Recovery
2) Rejection

In the application you are talking about rejection is normally the most important as your customers want de-ionized water. Recovery is important but you are talking about relatively small volumes overall. In single membrane system the only way you can achieve high recovery is by recycling some of the concentrate back to the feed. The rejection on a % of feed will stay about the same but the feed is a much higher salinity after a little while so your permeate will have a higher mineral content.
Depending upon the initial feed quality which will vary from site to site your recovery could be quite high (80% would not be unusual) but you would need an antiscalant, and higher operating pressures.

You are correct in your assertion that high pressures deliver better rejection, up to a point at least and higher flow for a given setup.

Assuming that rejection is your first priority and you want to stay away from using a an antiscalant, you probably need to do one of two things:
1) Trial permeate recycling. Feed some of the permeate back to the feed. This has the affect of reducing the mineral content of the feed water which at a certain % salt rejection will result in a lower conductivity permeate.
It also increases the feed and transmembrane flow rates and operating pressures but you will have still manage a quite high % concentrate flow. Because the recovery is still quite low you will probably not need an antiscalant.
2) Use a sea water membrane as this will have higher operating pressure and better salt rejection even at lower pressures and salinities.




Regards
Ashtree
"Any water can be made potable if you filter it through enough money"

 

[gerhard_za]

@ashtree,

Your last post has been very helpful! Thank you very much for that insight.

If you don't mind, please elaborate on the effect a permeate flush would have on scaling? I am considering small RO controllers for my systems which have the option to open an additional solenoid to run permeate water into the membrane after the system has shut down (i.e. permeate tank is full). Then only pure water will be in contact with the membrane when there is no flow.

I am looking at the cost benefit on adding a controller and reducing scaling. Do you think the reduction in scaling will be substantial for this option?

Second question I need to wrap my head around is whether using cation and anion resins in separate tanks would give me the exact same water purity as a mixed bed resin (provided the volumes of each resin is the same in both cases). This would help me to save costs by being able to regenerate both resins instead of discarding the entire bag after each round of depletion. I am yet to test an EDI setup but for now this is a bit expensive in terms of capital outlay.

Thank you all. Even though responses few on this platform, the quality of your inputs is immensely valuable.

Regards
Gerhard

 

[ashtree]

Gerhard,

It is normally good practice on shutdown to permeate flush the membranes so that they are not sitting there with various concentrations of mineral salts present. This will definitely increase the likelihood of scaling. Likewise many plants i am involved with permeate flush regularly anyway even if they are not specifically shutting down as this assists with the removal of small deposits of scale and particulate fouling as well as flushing out bugs, accumulations of biofoulants and anything else that might be on the concentrate side of the membrane.

There are a couple of points to note however.

1) Not all plants suffer from scaling. The formation of scale depends on the water quality, the temperature, antiscalant use, and recovery. A well designed and operated plant should not have regular episodes of scaling.
2) The pH of your permeate will impact on the efficiency of a permeate flush. A low pH permeate might be quite useful for dissolving small scale deposits but a pH above 7 will have limited to no effect.
3) The permeate flush will remove some particulate fouling etc just by the action of pumping through clean water.
4) To be most effective the permeate flush should not go from one stage to the next (if 2 or more stages are fitted) and there should be little concentrate back pressure applied.

Permeate flushing is a useful maintenance process that is effective on some plants but ineffective on others. Generally i would say its worth doing.


Regards
Ashtree
"Any water can be made potable if you filter it through enough money"

 

[gerhard_za]

Dear @ashtree

It has been a while since your last post, for which I am very thankful. I managed to find a simple solution to fill a pressure tank with permeate and have that line re-enter the source line with non returns. Then when the system shuts down the permeate flows out into the membrane and flushes the concentrate out, reaching about 10% of the original TDS. This will hopefully keep the membranes from scaling for a longer time period.

My next dilemma seems to be the presence of dissolved CO2 in the water after the RO step. Even though we get most of the ions out with the high rejection RO membrane, it seems as though CO2s flare up first and in some cases deplete resin within 3 months vs other sites that are able to run for 16 months!

Do you have any suggestions on how I might limit this CO2 to deplete the resin so soon. It might not be the CO2 specifically, it is just an assumption as that is what we pick up first in the reagent tests. Does CO2 pass through the RO more easily? Will it affect our permeate conductivity?

Thanks in advance.
Regards
Gerhard

 

[EdStainless]

What you are doing is balancing act.
You could choose a less selective RO system and let the IX do more of the work.
That costs more in regen, but less in running the RO.
We always ran separate Cat and An bottles, and in many cases followed by a mixed bed as backup.
If we saw one of the bottled depleted faster we could either go to a larger one or just parallel 2 of those.
Our alarms were after the single bed bottles.
The mixed bed was backup to give us a few days to get them changed without impacting output water.
We regened our mixed bed also, but it has to be done just right.
We had one application where we really needed low conductivity and we degassed the water.
Heated and bubbled argon through it, very coarse bubbles. What a PIA

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, consulting work welcomed

 

[ashtree]

CO2 formation in the RO permeate is largely dependent upon the pH of the feed water. A feedwater pH above about 8.3 will have little to no CO2 in its permeate in which the pH is likely to be 8.3 or above. As feedwater pH goes lower than 8.2 progressively more CO2 will be present in the permeate.

CO2 passes through RO membranes with little rejection. So almost all of the initial CO2 ends up in the permeate. Likewise most of the buffering elements have been removed by the RO process. Because all total dissolved solids, including alkalinity, has been removed, the permeate water has no buffer. With no buffer, the free CO2 lowers the pH to 5.0 – 6.0, depending on the concentration.

However simply increasing the feedwater pH to 8.2 may introduce other problems such as Calcium scaling. Such things are site dependent based on the water quality.

Regards
Ashtree
"Any water can be made potable if you filter it through enough money"

 

[gerhard_za]

Thanks Ashtree,
Your insights have been very valuable. I realise that CO2s might not be the root issue here as my resin supplier informs me that CO2 and Silica are the first elements that will be rejected and detected if the resin starts failing, so it is probably other elements depleting the resin and then I pick up CO2 first.

 

[bimr]

On large systems, the CO[sub]2[/sub]is removed using a degasifier. Not sure that it is practical for you.

Your best option is oversize the anion resin volume.

 

[dfmorvan]

Hi

Regarding CO2 depletion of mixed bed for a small system the easiest way is to use a softener instead of antiscalant and to dose caustic.
Other solutions exists (degassing membranes) but they will be over expensive.

 

[gerhard_za]

Thanks guys, I appreciate the efforts to answer my questions. I am learning a lot in the process.

I have learnt that I am in a precarious position with my application. I don't run the RO for long periods and it switches on and of numerous times per day. I am not in a position to use softeners as labs don't have space and we've tried it before - they won't maintain the salt refilling practice. Also, I've come to the conclusion that it isn't necessarily the gases that are the troublemakers. CO2 is merely the first element you pick up when doing calibrations for all reagents and that is only an indication that the resin is depleted - not that there is too much CO2 in the water...

So now I'm on the hunt to find out whether I should try and divert my first couple of litres of permeate before I send it to the resin. I have already installed a system that introduces the permeate back into the membrane after each use and it drops the brine TDS reading to 10% of what it was after shutoff. But still - when you start the system up, the it takes some time for the conductivity to settle at a nice and low value which is then obviously better for resin life.

Any comments or suggestions on this?
Thanks!

 

[dfmorvan]

Hi

your recyling alternative seems legit.