Nitrogen Tri-Chloride: The Explosive Contaminant
V K Kapur
Chloro-Alkali Consultant,
Mumbai

V K Kapur, a chlor-alkali industry veteran has narrated his experience of tackling contamination of liquid chlorine after the organization received complaints from its customers. The readers will be taken through an interesting trail of how the team finds the source of contamination, connects with the customers, and takes immediate steps to prevent the disaster.

We were in Caustic-Chlorine business for over three decades and had rarely received serious complaint about the quality of liquid Chlorine supplied to the customers. All of a sudden, we had three unusual complaints from our consumers within a short period.

Complaint no.1) A mild explosion was heard in a water treatment plant when the liquid chlorine ton container was nearing its end.

Complaint no.2) A small scale chemical factory reported Chlorine smell coming from their chlorine reactors event

Complaint no.3) A professionally managed company, M/s Arundhati Chemicals, reported bursting of the rupture disc of chlorine evaporators and more than 15 discs have exploded in last one . Their consumption was more than 1000 tons per month of liquid chlorine from the ton containers

First two complaints were investigated by sending area-in-charge but nothing unusual was reported by him, so no further action was initiated.

The last complaints of rupture discs bursting from a company managed by professionals and the earlier complain of the explosion sound in the water treatment plant, made us to think for a possible contamination in chlorine.

We were faintly aware that Nitrogen Tri Chloride (NCl3) can be the possible contaminant which is known for its explosive property.

Our factory had two cell houses .The first cell house was for membrane cells with its separate brine circuit to produce NaOH. The second cell house had mercury cells with facility to produce both NaOH and KOH; and had their separate brine streams. In total, we had three brine circuits. The chlorine gas produced in both the cell houses was cooled, dried separately, and liquefied to go to a common storage tank for ton containers filling for the customers.

Part of cooled chlorine gas before drying was diverted to produce Hydrochloric acid.

We were convinced that it had something to do with the NCl3 contamination only and contacted Chlorine Institute USA, & Euro- Chlor Brussels, asking them to send the available literature on NCl3 urgently.

In the meantime we analyzed various streams for presence of nitrogen compound and checked NaCl brines (two streams), KCl brine, Sulfuric acid (used for wet chlorine drying) and water (for the chemical preparation).

The NaCl brine solutions for the two cell houses (mercury cell and membrane cell), sulfuric acid and water; had traces of ammonia ruling out possibilities of NCl3 at such a low ammonia concentration. However, Potassium chloride brine had more than 50 ppm of ammonia.

The chlorine produced in the Potassium hydroxide (KOH) cells was 10 percent of total chlorine produced by both the membrane and mercury cell plants.

As we didn't have the facilities to analyze NCl3 in liquid chlorine, we advised our laboratory to analyze the chlorine gas of NaOH & KOH cell for NCl3 content and they asked for three days to set up the facility.

We studied the literature received from Chlorine Institute USA, & Euro-Chlor Brussels carefully which revealed following information:
  • 13 percent NCl3 in liquid chlorine appears to be limiting concentration to achieve detonation with change in the temperatures.
  • A mass of 1.5 grams of pure NCl3/cm2 on a metallic surface wetted with chlorine is capable of fracturing the metal of a typical chlorine ton container.
  • It has a vapor pressure much lower than the liquid chlorine.
  • Every 1 ppm of ammonia, present in the brine solution produces 50 ppm of NCl3 in chlorine gas.
  • For a safe plant operation 5 ppm ammonia in the brine solution is the limit as per the literature.
Our laboratory had reported 50 ppm ammonia in KCl brine stream ten times more than safe limits in a day. Based on the results, our team assumed the content of NCl3 at around 2500 ppm of NCl3 in the chlorine gas coming out of the KOH cells when KCl brine ammonia content is 50 ppm. After two days, the quality control laboratory submitted the report of gas analysis when the KCl brine had 35 ppm ammonia, KOH cell chlorine gas had 2150 ppm of NCl3, the mercury cell house (where the KOH cells were operating) had 710 ppm and the mixed chlorine gas of both the cell houses going for liquefaction reported as 09 ppm.

Naturally, we started analysis of KCl salt bags in the warehouse to know the extent of ammonia contamination which was found in the range of 500 ppm to 50 ppm that was too high as compared to the limit of 5ppm set for our supplier.

We advised the following steps for reducing the NCl3 formation:

A. To discontinue the use of KCl salt showing presence of ammonia and use old stock of salt lying with us from earlier supply
B. To inject open steam and air in all the three KCl reaction tanks to increase the temperature and provide vigorous agitation (higher temperature is needed for the removal of ammonia from brine solution and hot air drives away ammonia)
C. To increase the free chlorine content in return KCl brine by adjusting process condition, as free chlorine kills ammonia
D. To convert maximum possible KOH cells for NaOH production. We had the provision of converting few KOH cells to NaOH cells
E. To divert the chlorine gas of mercury cellhouse for hydrochloric acid production
After the implementation of steps A, B, and C , the level of ammonia reduced to less than 3 ppm in KCl brine.

Post implementation of the above mentioned steps, we started daily analysis of NCl3 in KOH cell house & the mixed chlorine that was going for the liquefaction. Our team found the results very satisfactory.

With 3 ppm ammonia in KCl brine, we were sure that NCl3 would be negligible in liquid chlorine going to the storage tanks.

Our visit to M/s Arundhati Chemicals:

As a next step we visited M/s Arundhati Chemicals armed with the literature received from Europe and the USA.

During the meeting we accepted the presence of nitrogen tri-chloride in our chlorine. We explained the presence of ammonia in KCl salt causing the formation of NCl3. We had admitted that we were ignorant about the source of contamination and assured to inform once we located the source.

As a precautionary measure we advised them to reduce the withdrawal rate of liquid chlorine from the containers to half and increase the temperature in the evaporator to 70-75 degree Centigrade. (Boiling point of NCl3 is71 degree Centigrade). In this way, the entire NCl3 dissolved in liquid chlorine will get vaporized and no NCl3 will accumulate in the evaporator's rupture disc.

We explained that using chlorine in liquid phase has very little chance of serious explosion compared to gaseous chlorine. When Chlorine gas is drawn from the top valve of the container, NCl3 doesn't get evaporated (due to low vapor pressure compared to chlorine gas) and gets concentrated in container itself, with high risk of explosion. However when liquid is drawn from lower valve of the container such situation can be avoided.

This container has two valves and when the container is positioned so as to bring the two valves in vertical position as shown in the picture below, the top valve gives the gas and lower gives the liquid chlorine.



We also advised our client M/s Arundhati Chemicals to use only 850 kg of liquid chlorine, and return the tonners with 50 kg of chlorine. We informed the steps taken for ensuring minimum NCl3 in future supplies. We had told them our inability to analyze NCl3 in liquid chlorine but gave gas analysis showing traces of NCl3 in mixed gas now. A copy of Chlorine Institute USA literature was handed over to them.

Protection for Gas Users from NCl3 Contamination : The next step was to protect the customers who were using chlorine gas only and we were scared about serious risk involved for them. We called for an urgent meeting of chlorine vendors and customers to inform them about some contamination problem that was resulting in choking of containers outlet valves. This was also reported by some of the consumers.

We advised them not to use total 900 kg of Chlorine but to draw only 850 kg and assured them to pay back for the 50 kg unused Chlorine. Further, we also offered to compensate the consumers who had to face the loss in production due to frequent choking of pipelines and outer valves.

At this point of time, we did not want to scare them with the information about contamination because of NCl3.

Locating the source of ammonia in KCl salt: KCl salt is imported and to make it free flowing, anti-caking agent is being added to it. We wrote to our Canadian supplier about the concentration of nitrogenous anti-caking agent used by them. But the supplier denied having used any ammonium compound in it, which ruled out the possibility of the presence of ammonia.

It is important to note that earlier KCl was being imported in HDPE bags that were packed at supplier's end. In order to reduce huge packing cost, we started procuring KCl salt in loose condition and bagging it at Indian port.

Now the only possibility that was left with us was that the consignment was contaminated at the Indian port. The consignment size of this shipment was 10,000 M.T that we had received six month back at Nhava Sheva port Bombay. We wanted to locate the source of ammonia contamination at Indian port now after a gap of almost 6 months which was similar to that of searching a needle in hay. Though we were not very hopeful of finding any clue, still we went there.

We found the jetty area to be very neat and clean where a 2 km long belt conveyor was used to unload loose materials from ship into any of the five warehouses used regularly for storage of food grains, fertilizer, and other raw material.

Our team collected various deposits found on the steel structures along the length of the belt conveyer from jetty to warehouses to examine the samples. The samples were muddy with a mix of wheat grains as well as several chemicals, but there was no odor of ammonia. We carried the samples to our lab for analysis. After collecting the samples, we went to the record keeping room at the port and found that the warehouse no 4 was exclusively used for chemicals where KCl that we procured from our suppliers, was unloaded and stored in the same process used for all other chemicals.

The size of the warehouse was approximately 15 m in height and 200 X 50 m with capacity of around 50,000 MT to store loose chemical. We found many small heaps of dirty material lying around and collected samples from different heaps, many of which had strong smell of ammonia.

Though this seemed like a breakthrough for our team, the first question that crossed our minds was: if ammonium salt was recently unloaded, we would still not be able to know how our consignment was contaminated. We further examined the eight months record of materials received in warehouse no 4 which confirmed that our consignment was unloaded just after the parcel of ammonium chloride after which no other chemical consignments were received in the same warehouse.

The port did not have proper system of cleaning the warehouse in place to prevent mixing of chemicals; and even requesting the port authorities for implementing that was of no help as they did not cooperate. Meanwhile, our next consignment was due to arrive at the same port in next 30 days and we took more than a week to wash and clean the warehouse before unloading of parcel. And as the consignment arrived, we got the same packed in our presence. After this experience, we received subsequent assignments at a port nearby Nhava Sheva and ensured carrying out proper inspection during unloading & packing of salt.

Handling of returned containers in factory: Post our meeting with our customers, we had already started receiving containers with 50 kg chlorine rich with unknown quantity of NCl3. We could not purge that for the reasons already explained.

According to the literature from Chlorine Institute USA, safe limit for NCl3 in ton containers is 20 ppm and 2 ppm in storage tanks; however 13 percent in NCl3 on steel surface of the container can result in an explosion. Since all the containers that were returned with 50 kg of chlorine rich with NCl3 were still intact which indicated the quantity of contaminant to be less than 13 percent.

Our team concluded that if we were to transfer the 50 kg of liquid chlorine parcels with less than 13 percent NCl3 to storage tanks, with 50 MT of chlorine capacity, the increase would be much less than 2 ppm and very much within safety limits. Soon after the detailed deliberations, we pushed chlorine into the storage tanks with the help of uniform mixing of dry air which was bubbled through one of the outlet dip pipeline.

After transfer of chlorine to storage tank, all the containers were air purged, washed with water, steamed, and dried. During the washing cycle, our team observed that the containers had more dirt, rust, and lot of brackish material as compared to containers that were emptied regularly for reasons unknown to us.

Utilisation of Contaminated KCl Salt
After all the exercise, we still had 500 MT of contaminated KCl purchased at the rate of around INR 12-13k per ton that was to be consumed somehow!!

We dissolved 5 MT per day of contaminated KCl in the sludge pit in chlorinated water with vigorous air agitation and open steam heating. Once the level of ammonia reached 5 ppm, it was pumped back to the system. It took us almost three months to exhaust the contaminated stock of KCl.

During this period, we received only one damaged container with its deformed welded portion of the end cover from a local consumer who did not report the accident at all.

It took us more than a year to overcome the NCl3 contamination in liquid chlorine. Had the bursting of the first rupture disc been reported immediately, we wouldn't have overlooked the issue and probably would have been able to take action much before.