Clever Denitrification Bugs; or How Poor Communication Nearly Derailed a Major Development

Article by Jimmy Hunter BSc CEng CSci

Modderfontein, Johannesburg, 1979

WE often see examples (on television and in the papers), of successful collaboration between professionals in the development of new technology. With the increased specialisation of professional people, such collaboration is even more important. The problem is that the different specialisms make communication between scientists and engineers from different fields even more difficult.

In most of the examples we see, the difficulties are glossed over. This story is one of a very interesting development which nearly went off the rails because of hubris and lack of communication between two professionals with different backgrounds.

At the time of this development, AECI had the biggest explosives business in the world. The original business was built on the production of dynamite for the gold-mining industry, but AECI had developed new products, using ammonium nitrate (AN), to produce explosives which were cheaper, safer and easier to use.

Although AN is water-soluble, scrubbing of AN fumes is difficult. Evaporation of AN solutions produces small particles which are charged, hence water-repellent in the vapour phase. Scrubbing systems produced large volumes of dilute AN which was difficult to dispose of.

The company tried some innovative solutions for disposing of the effluent. For example, since AN is rich in nitrogen, it is a natural fertiliser. AECI ran a cattle farm nearby, and the first idea was to spray the effluent as a nutrient for grass. However, AN is a very good nutrient for water plants as well, resulting in serious eutrophication of the river nearby. The other problem is that there was a limit to the ability of the grass to utilise all the AN, so the grass began to die. Another approach was needed.

Laboratory scale

We had to develop a process for economical removal of the nitrate from the effluent. The initial work was done in the research department, supervised by a PhD biochemist with substantial experience with biological processes. Our process development group, with our chemical engineering experience, were to take the laboratory-scale results and scale them up to a commercial scale in the form of a pilot plant.

This is the story of how technical arrogance and poor communication can easily derail an important development.

The biological process for nitrate removal used a carbonaceous substrate, such as methanol or molasses, as reducing agents for the nitrate. We can write:

CH3OH   +   3NH4NO3    --->  3N2    +   CO2    +   8H2O

Unfortunately, methanol is fairly expensive, so molasses was used as the carbon source. The equation for molasses (2{C}) can be written:

2NO3-    +   4H+   +   2{C}   --->   2H2O    +   2CO2     +   N2   

The sugars in the molasses provide the carbon source. The optimum pH is from 7.8 to 8.0. The reaction used denitrifying bacteria (“bugs”) to do the reduction. The volume of these bacteria steadily increased and could eventually cause blockages in the column.

Molasses: An inexpensive source of carbon

This may disturb certain gentle souls, but we found that we could reduce the feed of molasses so that the bacteria would become 'cannibals' and consume some of their own biological material as a source of COD. This prevented growth and hence blockages.

The laboratory-scale process used a glass column about 500 mm long and 150 mm in diameter, with glass Raschig ring packings inside. The denitrifying bacteria (our clever bugs) grew on the surface of the glass packings. The high-nitrate (>1,000 mg/L as nitrate-N) effluent was pumped through the columns and out to drain. Typical nitrate removal was 90%.

Design of pilot plant

We chose to use a polypropylene column and packing, which are much more robust and cheaper than glass. The column was about 1.2 m in diameter and 4 m long, and we designed the feed pump to give a similar residence time to that in the research column.

Source of organisms (bugs)

The normal source of active biological entities would be from a specialist supply house. However, the cost of enough organisms to coat the packings in the pilot plant would have been very high, and the delivery time would have been excessive. We decided instead to use sewage sludge. AECI operated a modern sewage plant, so we would be able to get sewage sludge as a starter material. The concept is that the sludge would contain a huge range of organisms, including denitrifying bacteria. As the effluent flowed through the sludge, non-denitrifying bacteria would die out, leaving the desired denitrifying species to multiply and take over.

People problems

We arranged for a trailer to be filled with the sludge and transported to the pilot plant. Our technicians would then manually load the sludge into the pilot plant column. This is where we hit our first “people” problem! The technicians refused to load the sewage sludge. They felt that such a job was beneath them. I tried pointing out (to no avail) that the sludge was not very smelly, and if they kitted up properly, with gloves and wellies, they would not contact the stuff.

Eventually, I put on overalls, gloves and wellies, and started loading. They soon felt ashamed and joined in. We got the material loaded in double-quick time.

Operation disappointment

We started up the pilot column and soon started getting denitrification, which was pleasing. However, we only achieved 45% denitrification, instead of the expected 90%. We waited a few more days, but the efficiency did not improve.

The biochemist said: “You must have made a mistake in your calculations.” I drew myself up to my full height and said: “Chemical engineers don’t make mistakes in their calculations”. Of course, I knew my calculations were correct, because we always used an independent checker to check all calculations.

We kept the plant running for another few days, but the efficiency did not improve. I asked the technician who had done the research work to let me see the laboratory notebook, and saw that they had added phosphate and magnesium salts to the feed to the research column. “What’s this about?” I asked the biochemist. “You didn’t tell us about these additives!” She said: ”Everybody knows that these organisms need phosphates and magnesium salts!”

The arrogance of the biochemist colleague in this remark was matched by my arrogance in assuming that I knew exactly how to scale up the process. I should have done more preparation by reading up about the mechanisms of biological denitrification, and had more in-depth discussion with her and her colleagues. Reading the laboratory notebooks is an essential activity before designing the pilot plant, not afterwards. My biggest mistake was that I had underestimated the complexity of the system. The other mistake was to allow personal feelings to cloud my judgement.

Understanding: Good communication with fellow professionals is essential

We immediately ordered a dosing pump and soon installed it, together with a feed tank for the additives. Within two days the nitrogen removal efficiency had improved to 90%. We ran the pilot plant for a period to generate data and make observations for designing a full-scale plant. One of the problems we found was that the bugs grew and started to block the effluent flow. We solved this problem by reducing the molasses flow, which resulted in developing “lean and hungry” bugs which no longer created a blockage in the column.

Once the column was working well, we decided to investigate a fluidised bed design for the process, using sand as the substrate on which the bugs would grow. We got this to work well, but had to solve a number of problems. The biggest problem was that nitrogen gas bubbles attached themselves to the bugs, making the bug-covered sand grains much lighter, so that they floated out of the reactor, resulting in sand and bugs losses from the system. We solved this by installing a spinning disc near the liquid outlet, to shear off the bugs and gas. The fluidised bed reactor increased the productivity of the process nearly 20-fold because of the improved mass transfer compared to the packed column design.

I left the company soon after this development, which was reported in the technical literature by some of my colleagues and the biochemist. However, it taught me a huge lesson in the importance of good communication with fellow professionals, and not to make assumptions. If the process had been a failure, this would have resulted in large financial losses and a major loss of prestige in our process development efforts.

Article by Jimmy Hunter BSc CEng CSci

Process engineering consultant, JHunter Process

Recent Editions

Catch up on the latest news, views and jobs from The Chemical Engineer. Below are the four latest issues. View a wider selection of the archive from within the Magazine section of this site.