Reverse Osmosis Membrane Production.
Factors that affect it.

David M. Bauman, Technical Ediitor of Water Technology magagine, answered a question that we frequently get about reverse osmosis membrane output in the magazine March 2007 issue.

We’re including the question and the answer as a reference to our customers.

Effects of temperature and pressure on reverse osmosis.

By David M. Bauman, Technical Editor

From the March 2007 edition of Water Technology magazine. Go here for a free introductory subscription.

Q : Since warm water produces more product with less waste, is there any reason that reverse osmosis (RO) units can’t be tied into the hot water or blended hot and cold? Raising the temperature from 55 degrees F (12.8°C) to about 70°F (21.1°C) would be a significant increase in production.

I’m just curious as to how temperature affects RO systems big and small.

Gallons per day (gpd) RO ratings are usually tested at 77°F (25°C) and 60 pounds per square inch (psi). But in actual practice the temperature and pressure can be very different. What happens if water temps reach 100°F (37.8°C) or more? Any damage to membranes? If I did this, the retention time in the RO tank would return water to ambient temperature before too long.

Would it tend to keep the membrane cleaner and lasting longer, as well? I suppose the easiest way to increase water production and quality is increased pressure before the membrane. Atmospheric tanks can offer greater permeate production than permanent air-head tanks, right?

I was just brainstorming on how to improve RO production. I have done some other research, and a couple of Internet sites say that 77°F (25°C) is the optimum temperature for RO production.

One unfortunate aspect of RO advertising is that the high gpd ratings shown are rarely achieved in real-life application. Customers are upset when a rating of 75 gpd is shown on the literature I give them, when they are only getting 15 gpd. All the factors that lead to this difference need to be explained or realistic numbers must be given.

A : I couldn’t agree with your last sentence more. You have asked some very good questions and have also raised an issue that I personally feel strongly about. Manufacturers and distributors are being irresponsible when they give you sales or technical literature with data that is practically useless to you.

Yes, there are some waters that are 77°F (25°C). They are in southern states, notably in Southern California, where much of the first membrane development was done. Hence, all the data was, and still is, presented with production figures at that temperature.

Why change it? It sounds pretty impressive, as you know, but it leads to big-time customer disappointment.

Correction factors
Here is a temperature correction table (sidebar) from a TFC (thin film composite) membrane manufacturer, modified slightly for ease of use. The factors are meant to be multiplied by the 77°F (25°C) published gpd to correct for temperature only. There will be other factors that might reduce the actual production. Note that 77°F (25°C) is not the optimum temperature but only the one that the membranes are tested at.

If you see a gpd production figure on literature at a specific pressure that doesn’t match yours, you can use a formula to make a correction. Note the meaning of the subscripts in the following formula: 1 = as shown in literature, 2 = actual gallons or pressure.

gpd₂ = (gpd₁ x psi₂) ÷ psi₁

You can also adjust for having a different total dissolved solids (TDS) reading than the one shown on literature. For every 100 parts per million (ppm) TDS you have above the TDS in the literature, you should subtract 1 psi from the “psi2” in the above formula. This adjusts for something called osmotic pressure, which works against higher production.

When customers complain
Here’s a customer complaint that should always be clarified before you even begin to address it: “We don’t get enough water.”

This can mean any of three different things and you need to find out which:

• It can mean that that the RO water, as delivered at its faucet, isn’t fast enough.
• It can mean that the RO tank doesn’t hold as much water as the customer expected.
• It can also mean that the RO tank isn’t being replenished fast enough.

The last two are somewhat related, meaning that improving one might make the complaint about the other go away.

Test case
To shed some light on these, here are a few numbers from my testing:

An RO tank 10 feet from the RO faucet delivered 1 gallon in 30 seconds. When the tank was moved to 30 feet away it delivered 1 gallon in 2 minutes and 48 seconds. Both of these included a 10-foot vertical rise from RO tank to faucet.

This is a significant difference for the customer drawing the water who wants to quickly get it into the fridge before they dash off to work.

Pressurized RO tanks apply a rising backpressure against the incoming water. This reduces production on a 1 psi-to-1 psi basis, just like decreasing inlet pressure. Therefore, it behooves you to find out if the production figure given to you by your supplier includes the tank backpressure. Or was it called “open flow,” meaning the water was being discharged to the atmosphere while being measured?

The tank will fill faster when empty and slower when it’s near full, due to the backpressure.

Averaging two “waters”
If there were no shut-off valve in the system, the pressure in the tank would eventually equal the incoming pressure and the quality of the last water produced would be of unacceptable quality and would degrade the quality of the first water. This last water would also be entering the tank very slowly.

From this you can see that the RO water in the tank was produced at a rate and quality that was an average of the first and last water produced.

When shut-off valves are used they prevent the tank’s backpressure from getting to the point where it seriously affects water quality; still, the last water to be produced is not quite as good and its production rate is not quite as high as the first water.

The degree to which this backpressure affects your production is dependent on tank size and shut-off point. ROs that discharge into a non-pressurized tank avoid this problem entirely but need a pump for repressurization.

Raising temperature
To improve TDS rejection by raising temperature is a good idea, but it is not advisable to use water from a heater unless your intended product water is not for drinking purposes.

You can, however, coil up some feedwater tubing to allow it to rise to room temperature, or you can wrap feedwater tubing around something warm, like the outside of a water heater. Your upper limit is somewhere around 95°F (35°C) before there is membrane damage. I don’t think you would improve membrane life by doing this, but you should check with the manufacturer.

As you suggest, the feedwater pressure can be boosted to increase both the quality and quantity of product water. I have seen RO systems that are suffering from low production simply because the feedwater line is too long or too small, both of which reduce pressure. Bigger is better.

Trying some adjustments
Let’s try the adjustments above on the example in your question.

You referred to a 75 gpd (from published literature) that actually produced about 15 gpd. After correcting for the 55°F (12.8°C) temperature, the production was down to 45.75 gpd. Use this for “gpd₁”.

I assumed your average pressure was 30 psi (as in a 20-40 psi well pump system) and then subtracted 4 psi for 400 TDS (assumed) and 5 psi for RO tank backpressure, making the psi₂ = 21 psi.

Using the “gpd₂” formula above, I calculated that your production would be 16 gpd, very close to your example. In the future you can calculate this in advance. However, remember this production is for a new RO that might suffer a little with age. This is still good production for drinking water in a home.