Matthew R. Simmons, Chairman of the energy-industry investment banking firm Simmons & Company International has written extensively on peak oil. He was interviewed on the subject by Foreign Policy Magazine back in 2005. And again by Energy Bulletin in June 2006

His firm has completed for its clients’ investment-banking projects that have valued over $65 billion. He has given 75 speeches since publishing his book on Saudi Oil, Twilight in the Desert (2005). “As I study the oil situation, the problems get worse… [but] the peak oil movement has grown from being a pimple to a pandemic,”

While the Peak Oil scare has died back a bit recently — Mr. Simmons has been quietly scaring the bejeebers out of Pentagon, DOE and Intelligence types for the last two years. (Mr. Simmons views were validated this week by a spate of news stories that reported the reason for Iran’s current production being below their OPEC quotas — was simply that they were running out of oil.)

Recently Mr. Simmons has decided to create “a new international water energy research center” in Rockland, Maine. According to Mr. Simmons

“What I’ve started is getting interested parties to get interested, hopefully, in Rockland, to create an institute in Rockland, an institute of water, and allowing 200 to 300 of the best scientists in the world, backed by maybe 20 universities, and 20 corporations and 20 think tanks, come here as a water fellow, and under one roof get all these people doing wave energy and tidal energy and desalination and so forth,” said Simmons.

Sounds like he could pull it off. But he could likely use some encouragement.

Certainly, I like any big idea that combines energy and water.

For several months I have been looking for work that might follow up on the Lawrence Livermore carbon nanotube break through annouced in May that promises in 5-10 years — to enable manufacturers to develop semipermiable membranes that cheaply desalinise water.

A couple weeks ago I posted a piece called Honey I shrunk the carbon nanotubes .

Recently some Japanese researchers figured out how to pack carbon nanotubes together effficiently.

Here are a couple more links to experimental work I’ve posted in the last several months that might follow up on the Livermore work. 1.) 2.) 3.) 4.) 5.) 6.) . This last #7.) discusses how charge could be used for semipermiable membranes.

The reason I mention this, is because according to reports widely published in provincial newspapers from Australia in October it seems that the Livermore team’s solution has already addressed all the problems for which I presented possible solutions. Did their solution come from some exotic process that’s not scaleable? Nope . They said they used “standard microfabrication techniques that in theory can be scaled.” Did the carbon nanotubes successfully filter out salt? Yup. They said current results showed that “nanotube membranes could remove up to 95 per cent of the chloride”–with better results in the offing. Was there any leakage around the sides of the carbon nanotubes? Nope. They said they deposited “a filler material – a matrix material that can fill the tiny gaps between the tubes to allow us to make a stable membrane.” Won’t these little nanotubes clog up quickly? Nope. The carbon nanotube “also showed that they have inherent anti-fouling characteristics.” Here’s the kicker. How come the larger carbon nanotubes were able to reject the smaller sodium ions. Guess. uh huh. Twas charge. “The fact that most of the sodium was not trapped by the experimental membrane pointed to the electrical charge mechanism being the active factor.” (There’s some serious implications here for hydrogen filtration as well.)

So why is this membrane 5-10 years away from a beta mass production model–and not, say, 18 months away?

Question 7How long do you think it will be before the technology you have been discussing becomes, if not everyday, at least usable?

Jason Holt – That’s a good question and one we often get. I think it depends on the kind of investment and partnering that we do with companies. There has been a lot of interest in the States from some of the big water treatment companies like GE and Culligan. They have contacted us and are interested in scale-up and commercialisation of the technology. If we get their involvement, a five-year time line will probably be overly optimistic – it’s more likely to be 10 years.

So it sounds like Mr. Holt is interested in working with smaller production companies that have already mastered the mass production of nanotubes with a a scaled up methodolgy simliar to his own methods for producing carbon nanotubes. Here are two possible candidates that I’ve pulled from back blogs that I’ve posted here and linked to above. One Nanovip.com and the other is NanoH20. There’s likely a couple more mentioned in my blog.

Finally a point should be made as to the real potential cost cutting here. According to the piece.

“The take-home message here is that the combination of fast flow and the small pore size suggests that we could use these membranes to significantly lower the energy cost of nanofiltration or desalination. A 100-fold higher permeability can reduce the energy costs by more than 80 per cent.”

If you just stuck a pipe in the ocean with a carbon nanotube filter on the end– the only enery costs would be pipeline pumping costs. Further, there wouldn’t be a big desal plant that would need to be built. What’s more the carbon nano tube membranes seem to have some anti fouling characteristics. So what the hey. Between collapsed energy, maintenance and capital costs — don’t you think perhaps 10 fold cost reductions can be done in 10 years– and maybe lot less?

Anyhow, this stuff is fun to consider. Here’s link to the brief article. In the meantime I’ve copied and pasted below a detailed article that covers the same event.

[Go to Home page] Australian Academy of Science | Conferences and lectures
Transcripts of lectures and speeches


Forthcoming conferences and lecturesConference proceedingsTranscripts of lectures and speeches PUBLIC LECTURE: Desalinating water cheaply – exploring technologies
Fast water transport through carbon nanotubes and implications for water treatment

The Shine Dome, Canberra, 26 October 2006 Dr Jason K Holt
Chemistry and Materials Science Directorate, Lawrence Livermore National Laboratory, California, USA


Jason HoltIntroduction (Professor Kurt Lambeck)I am pleased to introduce Dr Jason Holt who will speak on something that is very much on everybody’s mind – water! But what will be different from most of the recent water discussions, is that he does not see the solutions in terms of market forces and water trading, beloved by economists. He will be addressing new technologies that promise to produce more of the drinkable form through the use of nanotube membranes and desalination.

Dr Holt is a research scientist at the Lawrence Livermore Laboratory in California with experience in nanomaterials, nanoscale devices and analytical chemistry. His interests are very much in the application of this experience to alternative energies and the environmental sector, including nanotube membranes for water treatments.

Recently his research featured on the cover of Science and this brought his work to our attention. In the press release accompanying the article, he promised that a nanotube membrane on a silicon chip would offer a cheap solution for desalinisation. This is something that our coastal cities are most interested in and this is why Jason will be joining the Academy’s High Flyers Think Tank in Adelaide this coming week.

So without further ado, I invite you Jason to talk about Desalinating water Cheaply.


Jason HoltSlide 1
(Click on image for a larger version) Thank you very much for that introduction and thanks to all of you for coming here this evening. Thanks also to Sophia Dimitriadis for arranging my visit here.Tonight I will tell you about some of our recent work, our recent discovery of enhanced water transport through carbon nanotubes, and what implications that has for water treatment. What I hope to demonstrate today is that this is one of the areas in which nanotechnology can really benefit the environment. I know that there have been some concerns raised about potential environmental or health consequences of nanotechnology. I hope to show you that water treatment is an area where nanotechnology can really lend itself.I want to acknowledge some of my co-workers within our BioNanoSciences Group at the laboratory – in particular, Hyung Gyu Park who is a University of California Berkeley graduate student who has worked with me closely on this project – and also I want to acknowledge our two group leaders, Aleksandr Noy and Olgica Bakajin.Slide 2
(Click on image for a larger version) I will start off with some background information about the water shortage problem and present some relevant facts and figures that most of you will probably be familiar with.I will then introduce available desalination technologies, desalination being one of the technologies that can potentially address the water shortage issue.And then I will discuss our concept of a nanoengineered membrane, based on carbon nanotubes, which I think can address many of the shortcomings that conventional membranes, specifically reverse osmosis (RO) membranes, currently suffer from.

I will present some of our preliminary performance data showing how, in principle, implementing these nanoengineered membranes into an existing RO facility could significantly reduce the costs associated with water treatment.

Then I will wrap up my talk and show you where we are intending to go in the future.

Slide 3
(Click on image for a larger version)

So what is the problem that we are trying to address? This is nicely summarised by the three quotes on this slide. The general theme here is that to us globally, in the 21st century, water is going to be what oil was in the 20th century: a precious commodity. And the key thing to emphasise here when considering water versus oil is that there are a number of viable alternatives to oil but there really is no substitute for fresh water.

Slide 4
(Click on image for a larger version)

Our fresh water supplies have always been scarce – as you can see, just a very small fraction of the total amount of water on the planet is fresh water – and these supplies are dwindling. It is estimated that, if current projections hold, there will be a huge increase in the number of people around the world facing water scarcity, reaching as high as 18 per cent of the world’s population in the coming decades.

I also want to emphasise that 70 per cent of water usage is for agricultural purposes, often in very water poor areas – I think this is a global figure but the percentages probably look very similar in Australia. A case in point is China, the northern part of the country has two-thirds of the cropland but only has one-fifth of the water.

I myself am relatively new to the water treatment field and I am astounded by some of these statistics.

Slide 5
(Click on image for a larger version)

So then where should we go for additional water? You see here a list of the various alternatives, prioritised by relative cost.

One obvious choice is water re-use: recycling water from treatment plants and then potentially injecting that into aquifers for later use, or having a dual distribution where you separate out the non-potable water for use in irrigation and other applications where high purity water is not essential.

Next on the list is the treatment of marginally impaired waters. These are water sources that have maybe one or two species in solution, such as nitrate or arsenic, and none of the other impurities that you would have in, say, sea water. Some of the local water supplies in the Bay Area of San Francisco, where I live, are marginally impaired waters because they have high nitrate levels. Because of this, researchers within our Energy and Environment laboratory are developing so-called ‘smart’ membranes that can selectively remove nitrates from solution.

Finally there is desalination of sea water and brackish water, which has a salinity about one-tenth that of sea water. In the US, desalination of brackish water is more prevalent than desalination of sea water, although the US only uses a very small percentage of the total brackish water reserve.

Sea water desalination has not been widespread. This is because of the energy needed to drive water across the conventional low-permeability, reverse osmosis membranes. In fact, energy use is about 50 per cent of the cost of sea water desalination. If we can significantly reduce that energy figure we can potentially make sea water desalination more widespread.

Slide 8
(Click on image for a larger version)

We have here a list of the conventional desalination technologies that are available.

In reverse osmosis, a saline solution is pressurised, driven across a membrane – which prevents the sodium and chloride ions from permeating – and fresh water is produced.

Another membrane-based technology is electrodialysis. With this technique cation- and anion-selective membranes, along with electrodes, drive the positive and negative ions to opposite sides, producing demineralised permeate.

Another technique, which is commonly used in arid regions like north Africa and the Middle East, is thermal-based distillation, where heat is used to distil the water. Some of that heat is recaptured from vapour condensation.

Slide 9
(Click on image for a larger version)

So what do the energy figures look like? This graph shows the salinity, for both reverse osmosis and electrodialysis, plotted against the cost in megajoules per cubic centimetre.

Reverse osmosis, as I have mentioned, involves significant energy costs because of the high pressures needed to push the saline water across the membrane. The natural osmotic pressure has to be overcome, and the water needs to be driven across the membrane at significant flow rates: this is very energy intensive.

Electrodialysis suffers from similar problems. The membranes that are used are not very permeable to the ions; they have high resistivity. And electrodialysis is not very efficient except for brackish water. As you can see from the cost curves, at higher salinities, as represented by sea water, RO is the more cost effective alternative.

The thermal methods require significantly higher energy expenses. Theoretically, the minimum amount of energy required to separate water from the saline solution is about 1000 kW per acre foot. The amount of energy needed to distil saline water is 800 times more than that. This is why thermal distillation isn’t plotted on the curve. It is the most energy intensive of all the techniques and is really only suitable in regions like north Africa and the Middle East.

So, I would argue that of these technologies RO – because it can be used across a range of salinities – is the preferred option if we can improve the existing membrane technologies.

Slide 11
(Click on image for a larger version)

Listed here are the current problems that we face with reverse osmosis membranes, one of which is low permeability. I will show you later, with our concept of nanotube membranes, where we can have the greatest impact is in developing a high-permeability membrane.

Chemical degradation of conventional RO membranes is also a problem. The membranes are susceptible to fouling, which means that it is essential to pre-treat the water. Polyamide membranes can be degraded by chloride ions, which is a problem given that chloride is one of the major ionic species in sea water. Another type of polymer RO membrane, cellulose acetate, is susceptible to degradation when the pH is outside the target range.

I think nanoengineered membranes offer a potential solution to many of these problems, not just in permeability but also chemical compatibility.

So now I’ll tell you about carbon nanotubes and how we are using carbon nanotube membranes to tackle some of these problems.

Slide 12
(Click on image for a larger version)

For those of you who are not familiar with carbon nanotubes: nanotubes are atomically-smooth, molecule-sized channels. The tubes that we make are of the order of a nanometre (1 nm) in diameter. To offer a size perspective, 1nm is about 50,000 times thinner than a human hair and just a few water molecules in diameter.

The interesting thing is that water behaves very differently when it is confined within a carbon nanotube. This has been simulated for many years but not measured experimentally until very recently. Researchers predict that water forms unique structures within carbon nanotubes, and under the right conditions, it can actually form a one-dimensional chain of water molecules.

Perhaps because of the water ordering that takes place, it has long been predicted that the water would flow extremely fast through the channels. The best real world analogy we can come up with is that the water flow in these channels is like getting the same flow through a garden hose as you would get through a fire hose or a channel that is ten times larger.

On the bottom right-hand side of this slide you see an enhancement of the flow rate, of several orders of magnitude, through these channels. This is what had been predicted but it had not been measured until very recently.

Slide 13
(Click on image for a larger version)

How can we make a membrane out of carbon nanotubes? How can we take these interesting properties that the individual nanotubes possess and make a membrane that will allow us to test these predictions?

Our fabrication process is illustrated here schematically. I won’t go into all the details but I am happy to answer questions about the particulars at the end of the lecture.

In brief, we start out with a prepatterned silicon wafer. We deposit a thin bimetallic metal layer on the surface and heat it to a high temperature to form nanoparticle ‘seeds’. These serve as the catalytic seeds to nucleate the growth of carbon nanotubes. And because of the high density of these particles, the tubes that then grow from the nanoparticles are vertically aligned. They have a little bit of ‘squiggle’ in them, as is illustrated in this schematic, but the nanotubes do span from top to bottom.

The next step in the fabrication is to deposit a filler material – a matrix material that can fill the tiny gaps between the tubes to allow us to make a stable membrane. For that purpose we use silicon nitride (Si3N4). We have also demonstrated vapour-phase polymer deposition between the nanotubes, but for the membranes that I will talk about here we have used Si3N4 exclusively.

We coat the nanotubes, use a series of etching processes to open up the channels on either side, and the result is a nanotube membrane where the only pores in the structure are the carbon nanotubes themselves. The 2 x 2 cm square test chip that we made is shown here. I want to emphasise that although thetest chip is extremely small in scale, the techniques that we used are inherently scalable. We are using standard microfabrication techniques that in theory can be scaled up to a size that would be suitable for making a real reverse osmosis membrane.

Slide 14
(Click on image for a larger version)

I also want to show some microscopy to give you a sense of what the individual nanotubes look like and what the membrane itself looks like. These are transmission electron microscopy (TEM) images of carbon nanotubes. Image A shows bare carbon nanotubes and the inset shows an individual carbon nanotube with an inner channel diameter of less than 2 nm. It was essential to verify the size because the nanotube transport simulations that I mentioned earlier all focused on a regime less than 2 nm in size. Interesting physics happen at that scale and fortunately our TEM expert Yinmin (Morris) Wang went through the images to come up with a size distribution – it turns out that we have an average channel diameter of 1.6 nm.

Image C is a thin slice of the membrane, looking at it down the axis of the nanotubes. We wanted to verify that the only pores in the membrane were the nanotubes themselves and that there were no voids in the Si3N4, because we want to focus on transport through the nanotubes exclusively and ensure that there are no cracks in the structure. The bright spots correspond to holes in the membrane, and all the spots circled in yellow are individual nanotubes which have a channel diameter of less than 2 nm.

Slide 15
(Click on image for a larger version)

But the question is: how does the membrane perform in real life? What are its filtration characteristics? Can we actually filter out particles that are nominally larger than the 1.6 nm channel that we have measured?

So we carried out a series of filtration experiments where the membrane is mounted in a flow cell and placed in a solution of nanoparticles (the nanoparticles are upstream of the membrane). We apply pressure and collect what comes out.

We started out with a solution containing 5 nm gold particles – these particles are about three or four times larger than the pore size – and we saw no evidence that these particles permeated through the structure. This gave us a good indication that there are no large cracks or voids in the membrane.

We then tried smaller particles, at 2 nm – just nominally larger than the nanotube pore size – and again we saw no evidence for passage of these particles through the membrane.

The only species that we saw permeate through the membrane was a 1.3 nm organometallic dye. Again, this gave us a good indication that the membrane is indeed void-free.

As a control, we also looked at a multiwall carbon nanotube membrane. This has a much larger pore size, about 6 to 7 nm in diameter. It also showed filtration characteristics consistent with the measured pore size; allowing 5 nm particles to permeate but not 10 nm gold metal particles.

Slide 17
(Click on image for a larger version)

After confirming that the membrane was void-free, we wanted to tackle the physics question: what is the flow rate through an individual carbon nanotube? We needed to answer this so we could compare it to the simulation predictions that had been made several years ago. We also wanted to find out how the nanotube pores compare with those of conventional membranes.

We looked at a series of three carbon nanotube membranes, as well as a polycarbonate membrane – a conventional membrane with a pore size about 10 times larger than our nanotubes and the smallest pore size commercially available. I should emphasise, that to quantify the flow rate per nanotube we need to know how many pores are actually open and expanding the membrane structure and being wetted with water. As you might imagine, this is difficult to pinpoint exactly, but we can come up with an upper limit by using the microscopy images that I showed you earlier.

So, going through and counting the number of pores in several different regions, we come up with an extremely high pore density at 250 billion pores per square centimetre. If we use this pore density as our upper limit, it will allow us to come up with a minimum flow rate per nanotube.

We found that the flow rate per nanotube is orders of magnitude higher than what conventional hydrodynamic theory would suggest. By comparison, the polycarbonate membrane that we measured showed only a very slight enhancement.

Another way of trying to capture this is in terms of what is called slip length – I won’t go into the theory or the equations, but the slip length is an indication of how frictionless the flow is: large values would indicate that the flow through the nanotubes is nearly frictionless. This is something that simulations have long predicted, but it is a matter of being able to measure it experimentally on this scale – less than 2 nm. The flow rates that we measured are comparable with what the simulations have long predicted.

Slide 18
(Click on image for a larger version)

What does the physics of the flow rates mean in practical terms, particularly for membrane developments and desalination?

I have plotted the permeability for the nanotube membranes and compared it with the permeability for conventional polycarbonate membranes. Permeability is what people in the membrane community care about – that is, the volumetric flow rate per unit area, divided by the pressure drop.

As you might expect from the enhanced flow seen through the tubes and the high packing density, we see a significant enhancement in water permeability when compared to conventional membranes – in fact, about a 100-fold higher water transport rate.

As an aside: we also see enhanced flow of gases. I haven’t discussed this because the application that I am focusing on is water treatment, but if anyone is interested in discussing that afterwards I would be glad to.

The take-home message here is: the combination of fast flow and the small pore size suggests that we could use these membranes to significantly lower the energy cost of filtration – nanofiltration or desalination, in this case.

Slide 21
(Click on image for a larger version)

So how would this carbon nanotube membrane perform in reverse osmosis applications? We worked with systems modellers to ask the question: what effect would this higher-permeability membrane have on a reverse osmosis facility? The schematic of the model is shown here. We consider the feed stream as well as the permeate stream, and then the brine or concentrated stream.

One thing I should mention is that, unlike a real facility where energy recovery takes place in the high-pressure brine stream, we don’t take this into account in the model. But, in terms of showing how the energy costs scale with permeability, the inclusion of the energy recovery is not really important.

This table shows the impact of enhanced permeability on energy costs. For a typical polymer RO membrane permeability is 10-11 m3/m2-sec-Pa. Our nanotube membranes, as I mentioned, are about 100-fold higher. The cost which is shown here is the energy cost of desalinating, in terms of kilowatt/hours per cubic metre. We include here not only the energy costs of operating the pumps to produce the required pressure but also the energy to pump in the water from the source. You can see that a 100-fold higher permeability can reduce the energy costs by more than 80 per cent. So there is potentially a very significant impact on the energy costs of treating sea water.

In this model we fixed the flow rate, but what if you keep the pressure the same, what effect will higher permeability have on energy cost? Another way of looking at the problem is to ask: if we operate at the same energy or same pressure, what effect would that have? And, as you might imagine, from 100-fold higher permeability at the same pressure we would get 100-fold higher flow rate of water through the membrane. Which way a given plant should operate is perhaps not such a straightforward optimisation problem and may depend on the particular peculiarities of that region. But it shows you that, either way, with the high-permeability membrane we can, in principle, have a large impact.

Slide 22
(Click on image for a larger version)

What does this mean in dollar terms? A way to look at this is the permeate costs per cubic metre. Here we have included both the cost of the membrane and electricity. On the left hand graph we show permeate costs versus membrane area for three curves in order of increasing permeability and we find a minimum in the permeate cost at about 8000 m2. In the graph on the right hand side, we took that 8000 m2 value and applied it to cost versus permeability and we see a significant reduction in the cost – from about $1 per cubic metre down to 40 cents.

So, in addition to the significant reduction in energy costs, we can reduce the actual dollar cost of treating a given volume of water by about 60 per cent.

Slide 23
(Click on image for a larger version)

The other question is: what kind of selectivity does this membrane exhibit? It is great to have fast water flow through the tubes, but if we can’t reject salt to any significant level then it is obviously not very useful for desalination.

These are very preliminary data that we obtained just in the last month, but our initial findings are that the membranes show very encouraging salt rejection levels. It is important to note that for these experiments we only do a single pass through the membrane. A conventional RO membrane which you might be familiar with is essentially a spiral-wound membrane around the feed tube, so the water has multiple chances to pass through the membrane structure. In effect, it is like having many membranes in series so in a single pass almost all of the chloride is rejected.

With our membrane, the rejection of sodium is slightly less but, in principle, if we had a number of these membranes in series or, if we formatted it as shown on the right-hand side of this slide, the sodium rejection levels could approach 90 or 95 per cent.

What we don’t know at present is the mechanism: why is the membrane behaving in this fashion? Is there physical size exclusion of these ions from the pores, or is it due to surface charge effects? The initial results suggest surface charge effects dominate, but other experiments that we have performed under slightly different conditions show different rejection characteristics. These results are still very preliminary.

The nanotube membranes have some other properties that I would like to point out. They seem to have inherent antifouling characteristics. They don’t readily clog, and when they do we can regain the original permeability just by briefly washing the membrane with deionised water. There are additional parametric studies we would like to carry out, to see how the rejection characteristics vary with the size of the ion and the effect of pH and ionic strength.

Slide 24
(Click on image for a larger version)

To conclude: reverse osmosis desalination has long been touted as one of the viable solutions for our water scarcity problem, but there have really been no major developments in membrane technology for decades. I think that is what has made sea water RO economically infeasible. In fact, in the US there is really only one large-scale sea water RO facility that has been built, and it has been tied up in a series of legal and technical problems. In principle, it is supposed to come online later this year. The fact that we only have one such facility in the US highlights that operation of these plants is not economically feasible.

I would offer that the carbon nanotube membranes that we have developed show unprecedented water permeabilities and if they can be scaled up we can significantly reduce energy costs associated with reverse osmosis.

So we are taking a two-pronged approach here. One addresses the issue of scale-up: can we make these membranes on an industrially relevant scale? I think the answer to that is yes because we are using conventional fabrication techniques that are amenable to scale-up.

The other approach is to also tackle some of the science questions. We recently started a project to look at some of these issues, and the main question that we want to tackle is: why are we seeing these transport rate enhancements in nanotubes? Is it a consequence of structural ordering that occurs within the tubes? We don’t know yet, but we will use experimental X-ray based techniques to hopefully answer some of these questions. The science we are doing here is not just for science’s sake but because we think that by answering these questions we can actually design better membranes.

Slide 25
(Click on image for a larger version)

I would like to acknowledge the intellectual and technical contributions from the people listed here, as well as the three year internal funding we received from our Laboratory Directed Research and Development Fund to get this project to the stage it is at right now.

Thank you for your time, and I would be happy to answer any questions.


Discussion Kurt LambeckThanks very much, Jason, for a very stimulating and informative talk. Question 1Jason, could you go back to your slide entitled ‘What kind of selectivity does a CNT membrane exhibit?’, where you discussed the size of the membrane. You mentioned there that a water molecule was 0.3 nm in diameter. The CNTs were 1.6 nm. That means that the water would go through. But it seems to me that not only the water but also the sodium ions and the chloride ions would go through. The nanotube is much bigger than any of those.Slide 23
(Click on image for a larger version) Jason Holt – Yes, that’s true. If the membrane was working on the basis of size exclusion alone, we would in principle see no preferential rejection of chloride. But that’s why I made mention of surface charge effects. What may be happening is preferential rejection of ions that are the same charge as the surface of the membrane. That is perhaps the reason why we see such a high level of rejection of chloride relative to sodium.Question 1 (cont.)So the sodium ions and the chloride ions remain on the surface of the membrane. Don’t they clog it up? Jason Holt – No. On the figures that I mention here, we frequently see preferential rejection of the chloride – most of it still remains in the feed stream, but we only see a small rejection of sodium. So the sodium actually, by and large, can pass through. But over the course of these experiments we haven’t seen significant clogging of the membrane. It seems reasonable to expect some accumulation of those ions near the surface of the pores but we don’t see any reduction in the flow rate, at least over the duration of these experiments that we have carried out.Another thing to mention here, in terms of practical use of this membrane, is that ordinarily you would decide that there is cross-flow occurring, which would really help to prevent accumulation of ions near the surface. We are doing these experiments on the research scale with just a single pass of the membrane, but in practice you would have cross-flow. There are other engineering designs that you can use to minimise the effect of ion concentration near the surface of the membrane.Question 1 (cont.)The permeate then would have a high sodium ion level?

Jason Holt – Yes. In these initial experiments so far we see high sodium content but a low concentration of chloride.

Question 1 (cont.)Is it possible to remove that sodium concentration?

Jason Holt – Yes. I think one way of doing that is to change the surface charge on the membrane. So, instead of the membrane having the negative charges shown on the slide, there are simple chemical routes to converting those to positively charged groups. So in practice what we might have is a multilayered membrane: one like the electrodialysis membranes that I mentioned before, where you have alternating cation- and anion-selective membranes so you can achieve rejection of both species.

Question 2There has been criticism that the salt by-product will be discharged back into the sea, with environmental effects. Have you had any thought about that process, or is there any way of preventing that salt from being discharged back into the sea?

Jason Holt – As far as I know, the Tampa Bay ecosystem – where the only large-scale sea water RO plant in the US is located – appears to allow for it. But the brine is not being discharged directly. As I understand it, the brine is being diluted by a factor of 50 and then discharged into the Gulf of Mexico. But the environment is something that has to be considered wherever a plant is going to be operated. I am not aware myself of any alternatives to discharge, but I know that is a concern regarding the sea water at really concentrated high-salinity RO facilities.

Question 2 (cont.)I have another question. You opened with the scarcity of fresh water, and how it is disappearing. There is a school of thought that the amount of fresh water we received at the time when the earth was formed is what we always have, and our use of it doesn’t diminish that supply very much. It is a question of that quantity of fresh water being unevenly distributed around the world: we are not getting any here in Australia while someone else is getting our share. But basically the volume of fresh water remains constant.

This is not your area, I suspect, but I am just wondering whether anyone has any comments about that, and whether or not that in fact is proven – or thought – to be the case.

Jason Holt – I don’t have any insights myself. Does the question relate to the distribution of fresh water?

Question 2 (cont.)Well, it is not the question of quantity but the question of distribution that we need to tackle. I don’t know how factual that is.

Kurt LambeckThe total water is more or less conserved but there is no reason why the fresh water should be conserved. Take a case where we just make a mess of all the fresh water we have. I can think of no reason for all of that to be conserved.

Question 3Jason, you made mention of the use of Si3N4 as the filler material in the carbon nanotube forests. I am aware of other people using organic-based materials, rather than the material that you use. Would you care to comment on the pros and cons of Si3N4?

Jason Holt – We started using Si3N4 because that was the deposition process available to us. And it turns out that Si3N4 actually forms a very conformal coating around the nanotubes. I don’t have the SEM image to show this, but it does a nice job of filling the gaps between the tubes, which is really the key to making it a stable membrane. It is true that for making RO membranes we wanted to go the polymer route, and I had done some preliminary work looking at vapour-phase deposition of polyimide and parylene. It turns out parylene might be the best choice. It is very resistant to chemical degradation and you can actually reflow parylene at extremely high temperatures, up to 300 or 350°C. So that might be the route that we take in the future – use the same process but with polymer deposition and, essentially, make a flexible nanotube membrane.

Question 4Do you need the matrix material? Could you pack the tubes together to allow the water to flow around the outside of the tubes?

Jason Holt – It’s a good question. Our feeling is that the flow rate enhancements that we see are unique to the interior of the nanotube: flow through the interior of the nanotube is not necessarily going to be the same as the flow between a bundle of nanotubes. That’s part of the answer.

The other, I think, is an issue with the fragility of the membrane. Even a densely packed bundle of nanotubes with no filler material in between them is extremely fragile. You can easily tear a densely packed bundle of nanotubes. So for practical reasons I think we have to use a matrix material approach.

Question 5Just following on from those last two questions about the membrane: I was wondering, as a consumer of desalinated water, what might happen if I swallowed a bit of your membrane. Would there be a risk from ingesting a nanotube?

Jason Holt – That’s another good question. I think a lot of the health concerns about nanotubes, and other nanoparticles for that matter are about free nanoparticles or nanomaterials in solution – and in the United States there are a number of projects looking into health effects in nanotechnology. With this structure here, in particular the nanotubes are firmly embedded in the membrane; they are not free to move. Unless we are working at extreme pressures where we can actually fracture the membrane – which we are not currently doing and it’s not how we would be operating these membranes in a plant – I think there’s minimal concern with respect to the nanotubes becoming dislodged and getting into the water supply.

Question 6Jason, you’ve got the membrane there; on a single pass you’ve got the negative surface charge, so presumably that would retain the sodium; and then you’d have to have the second pass, which would be positive charge and would repel the chloride ion. How do you put these charges on the membrane? You would have to have some dialysis system, or how do you do that?

Jason Holt – This would be done by changing the surface chemistry. We don’t know this for certain, but most likely the negative surface charge on the membrane is due the presence of acid groups – you have carboxylic groups on there, COO- – and with that kind of surface charge there are easy chemical routes to switch that over to a positively charged group, something like an amine group or a basic group. Although we haven’t done that yet, I don’t think it’s a huge barrier to alter the surface chemistry to make a membrane that will reject cations.

Question 6 (cont.)So you have to have two passes?

Jason Holt – Yes, that’s what I think. I think that in practice, unless you can make a membrane with pores that are so small that you’ll be able to reject both species just on the basis of size, it’s going to be hard to make a membrane that will achieve similar rejection levels for both positive and negative ions. I think that practically speaking it may be easier to adopt a multi-layered approach.

Question 7How long do you think it will be before the technology you have been discussing becomes, if not everyday, at least usable?

Jason Holt – That’s a good question and one we often get. I think it depends on the kind of investment and partnering that we do with companies. There has been a lot of interest in the States from some of the big water treatment companies like GE and Culligan. They have contacted us and are interested in scale-up and commercialisation of the technology. If we get their involvement, a five-year time line will probably be overly optimistic – it’s more likely to be 10 years.

Question 8If I might just briefly get to the gas filtration properties of your membranes: I read in your paper in Science that there was a differential selectivity between carbohydrate and non-carbohydrate gases. I wasn’t exactly sure whether that meant that the carbohydrate gases were preferentially passing through the membrane or whether it was the other way around.

Jason Holt – In addition to higher transport rates for all gases, we also see an enhancement of the overall flow rate, with preferential diffusion of hydrocarbons, if you look at C1 through C5 – methane, ethane, all the way up to pentane. We actually see an enhancement in the transport of those gases, compared to what conventional diffusion theory would predict. And it kind of makes sense, based on what is out there in the simulation literature. One idea is that, because of the similarity in structure between a hydrocarbon molecule and the nanotube surface, it may be that these gases actually adsorb onto the nanotube surface and then can diffuse along the surface, as opposed to simply bouncing off the nanotube walls, so to speak. In addition to the water experiments we are also doing some more detailed measurements to try to better understand the gas transport mechanism, because at this stage it is still speculation as to why you see the enhancement in hydrocarbon diffusion.

Question 8 (cont.)It could also be that it has got more intimate interaction with the carbon nanotube inside. It seems that it is slipping through more easily.

Jason Holt – Yes, that is a possibility.

Question 9 I am just wondering what awareness you have of interest or understanding by water supply authorities or industry in Australia in relation to this technology. What sort of understanding or interest is there in Australia?

Jason Holt – We haven’t been contacted, at least in the past few months, by anyone regarding the technology, although we had a discussion today with Ken Matthews from the National Water Commission and he seemed very interested in the technology.

Question 10Jason, I gather from some of the acronyms in your earlier slides that these are double-wall nanotubes. How tight is the control over that? Are they all double-wall, or do you have some single and some multiwall mixed in?

Jason Holt – I glossed over that subtlety because for the purposes of making the membrane all we really care about is the innermost tube. But in reality, the tubes that we are making are, essentially, concentric single-wall tubes. From what we have seen in TEM – of course, it’s hard to get statistics on these things – we have counted about 500 of these tubes and looked at different regions in the TEM, and they all appear to be double-wall tubes with the size distribution that I showed earlier – about 1.5 nm inner diameter and the size cut-off is at about 2 nm. We don’t see any diameters larger than that.

The way we can control that is by tailoring the thickness of the catalyst layer. We are using, as I mentioned in the paper, an iron and molybdenum catalyst layer supported on aluminum, and by adjusting the thicknesses of these catalyst layers we can in turn affect the size distribution of the nanotubes. In fact, just recently we have been able to make similarly structured single-wall nanotubes – we found a recipe that allows us to make this same kind of membrane but comprising single-wall tubes, also vertically aligned and have a slightly smaller pore size – 1.3 to 1.4 nm instead of 1.6 nm.

Kurt LambeckThank you very much, Jason. This has been a very nice example of how basic science can be put to good practical use. I wish some of our masters were present to hear that. Thank you very much.

Jason Holt – Thank you.


[ Home | Contacts | Publications | Search ]


© Australian Academy of Science

GE Global Research and Texas Tech University announced this week that they are teaming up to work on another kind of desalination/energy combo.

GE to Develop Wind-Powered Water Purification

November 29, 2006

NISKAYUNA, NY — GE Global Research, the centralized research organization of the General Electric Company, today announced it is partnering with Texas Tech University to develop affordable water desalination systems to increase the quantity and quality of clean water available in arid areas around the United States and globally.

The GE-Texas Tech partnership will focus on the integration of renewable energy systems, such as wind turbines, with membrane desalination processes. The development of the integrated renewable energy-water system has the potential to significantly reduce the cost of creating new sources of freshwater from impaired resources, such as brackish water, by directly addressing the major component of operating cost of desalination systems – energy.

Again, notice that while there’s not much money available for desalination research–there’s plenty available for energy research.

The partnership is part of GE’s company-wide ecomagination initiative, in which GE has pledged to more than double its level of investment in the development of cleaner energy technologies, from $700 million to $1.5 billion, over the next five years.

btw why did GE choose Texas and why wind energy?

The answer is that Texas has recently become the largest wind energy producer in the USA with more soon to come online:

Texas has mandated that at least 5,880 megawatts of energy used in the state come from renewable-energy sources by 2009 and 5,000 more megawatts by 2015.

By the numbers

2,700 Megawatts of wind energy that could be added to the Texas grid by the end of 2007.

2,631 Megawatts of wind energy currently operating in Texas.

2,044 Turbines currently in Texas.

Top wind-energy states

1. Texas

2. California

3. Iowa

SOURCES: ERCOT, American Wind Energy Association

Wind power is expected in 2006 to provide 18% to 20% of the new capacity installed in the country — making it the second-largest source of new power generation after new natural gas plants according to the Energy Information Administration.

wind energy is one of the most economical forms of utility-scale renewable energy available, with a “bus bar” price (which does not include transmission and distribution costs) of 3 to 6 cents per kWh at good wind sites.

By contrast, power generated by oil-, gas- and coal-powered plants feeding into the PJM Interconnection — the grid operator covering most of the country from the Hudson River to the Chicago area and as far south as North Carolina — costs 2 to 3 cents a kilowatt-hour, the report said. PJM supplies power to 51 million customers.

Also, as I’ve mentioned before west Texas has a lot of brine aquifers below the windy desert that could see some benefit from desalination. These aquifers are currently being tapped by gas drillers. In some places the water used for drilling is being recycled and desalinated. These might also be good places for greenhouses.