Controlling Nitrogen Oxide (NO x ) Emissions from Exothermic Nitrogen Generation Systems for Application in Subsea Environments.

A highly exothermic nitrogen generation system (NGS) can be achieved by mixing solutions of sodium nitrite and ammonium chloride, a process used by the oil and gas industry to dissolve paraffin wax and gas hydrates. Although its main products are nitrogen gas and a sodium chloride brine, the NGS has a side reaction that produces nitrogen oxides. To optimize this process to ensure the greatest and fastest heat generation with the lowest oxide production, this reaction was checked by infrared spectroscopy and calorimetry. The factors temperature, pH, and initial concentration of nitrite and ammonium were evaluated, and the optimal conditions of the NGS were determined by the constructed models to predict heat and NO x generation. These conditions were a ratio of ammonium/nitrite equal to 1 and a catalyst concentration of 0.07 mol·L-1 (for a case in which the temperature is 5 °C).

Introduction

The reaction of ammonium and nitrite ions produces a highly exothermic in situ nitrogen generation system (NGS).[1−3] When ammonium chloride and sodium nitrite are used, this reaction is considered “green” because nitrogen and sodium chloride are the main products of the reaction (eq ).

Because of the high heat generation, this reaction has been successfully used as a method of fluidization of low melting point deposits in the oil and gas industry.[3−8] These deposits may cause a decrease in the oil flow or even blockage of production lines.[9,10]

In wells in the Gulf of Mexico, which produce oil with a high paraffin content, it was possible to increase productivity by up to 10 times with this method, a result that was not achieved with other remediation methods, such as the use of solvents.[1] In Brazil, the technique has also been used several times, such as in dewaxing operations in the Campos Basin,[11] melting gas hydrates,[7] or even dispersion and fluidizing different types of organic deposits present in crude oil storage tanks.[8,12]

One application of the NGS was to eliminate a hydrate plug formed in a subsea equipment that prevented it from functioning properly. The heat generated around the body of this equipment was able to increase its temperature and eliminate the hydrate.[12−14] Because NGS fluids were discharged to the seabed and to the atmosphere, it is essential to assess the environmental aspects associated with the use of this technique.

The NGS reaction begins as soon as the solutions of sodium nitrite and ammonium chloride are mixed. However, it is possible to control the reaction rate by manipulating the pH of the medium: a solution at pH 8.0 reacts very slowly compared to the other at pH 5.0.[1] For several reasons, a rapid increase in the temperature is desirable, so the pH decrease is mandatory.

The reduction of pH, aimed at accelerating the NGS reaction, increases the occurrence of an undesirable side reaction, which is the production of nitrogen oxide, NO = NO2 + NO (eqs and 3). Nitrite solutions can produce NO at pH values below 6.0, and the kinetic of NO becomes even faster when the pH drops.[15]

NO has numerous negative impacts on human health and environmental balance (air quality, climate, and ecosystem health).[16] Although NO is less harmful than NO2, nitrogen monoxide is unstable and reacts readily with oxygen to form NO2 (eq ), which is dangerous even at low concentrations.[17] Exposure to nitrogen dioxide causes irritation and may cause a predisposition to respiratory disease, weakening bronchopulmonary structures and triggering acute lung injury.[17,18]

Because of the possibility of conversion to acids (eqs and 6), NO is capable of greatly disrupting aquatic ecosystems and may cause the biological death of water bodies, which can also undergo eutrophication because of increased availability of nitrogen.[19] In turn, when these oxides reach the atmosphere, they cause, directly or indirectly, environmental problems such as photochemical smog, acid rain, and even global warming.[20−25]

By considering the environmental issues related to NO and the lack of studies on the balance between heat and NO production by the NGS process, this work aims to perform a quantitative analysis of the NO production and evaluate the most appropriate conditions to ensure the greatest heat generation with the lowest oxide production.

The real-time monitoring of the NGS reaction,[26] performed by measuring the nitrite and ammonium consumptions using attenuated total reflection Fourier transform infrared spectroscopy (ATR–FTIR), was used to evaluate the NO generation. If no side reaction occurs, the ammonium concentration should decrease at the same rate as the nitrite consumption. Therefore, the comparison of the in-line monitoring of the reaction progress by evaluating both nitrite and ammonium ions could shed light on the understanding of how nitrogen oxide generation can be lowered in the NGS process.

Materials and Methods

Materials

Solutions of sodium nitrite and ammonium chloride were used in the concentrations of 8 and 4 mol L–1, respectively, prepared with ultrapure water (type I) of a Milli-Q apparatus (Millipore, Burlington, MA, USA). Glacial acetic acid was used as a catalyst, and ammonium hydroxide was used as the NO absorbing solution. All reagents were purchased from Sigma-Aldrich (St. Louis, MO, USA) and had an analytical grade.

Equipment

The experiments were carried out in a Mettler Toledo RC1e calorimeter. The reactor connected to the calorimeter was an HP60 double-walled vessel with a 1.8 L volume, in which the temperature of the reactor contents was controlled by varying the temperature of the fluid circulating in the vessel jacket. Several accessories were attached in the reactor cover: a stirrer shaft with a propeller stirrer, a Pt100 temperature sensor, and a 15 W calibration heater. The reaction vessel and all components were produced in Hastelloy C22.

The real-time FTIR monitoring was performed by a Mettler Toledo ReactIR 45m spectrometer, whose base unit contained the Fourier transform mid-infrared source and a mercuric cadmium telluride detector. A K6 (16 mm diameter) conduit contained the optics that transfer the infrared light source from the base unit to the probe in contact with the reaction mass. Measurements were taken optically using a diamond sensor as the ATR element, where the IR-beam is attenuated if there are regions of the IR spectrum where the sample absorbs energy.

Method

Specific volumes of the standard solutions of sodium nitrite and ammonium chloride were mixed in the HP60 vessel so that the ratio between the final concentrations of these species, [NH4+]/[NO2–], was as defined for each experiment, always totaling 0.6 L final volume (Table ). Thereafter, the temperature of the medium and stirring (350 rpm) was maintained at a constant level throughout the experiments.

Table 1

Volumes of the Standard Solutions Employed in the Preparation of the NGS Reactions

[NH4+]/[NO2–]	ammonium chloride (mL)	sodium nitrite (mL)
0.50	300	300
0.75	360	240
1.00	400	200
1.25	429	171
1.50	450	150

The ATR–FTIR spectra were collected in the range of 4000–650 cm–1 with 8 cm–1 resolution during all the reactions. The beginning of the reaction was considered the moment when the catalyst was added. Finally, a specific volume of the catalyst was added to the medium so that its final concentration was as defined for each experiment. Data were collected for 11 h after the catalyst addition.

In the direct quantification of NO, these oxides were absorbed in 100 mL of a 6 mol L–1 ammonium hydroxide solution during the NGS reaction. Afterward, excess of sodium hydroxide was added to this solution, which was heated until the generation of ammonia vapors ended. Next, the acidity of the solution was adjusted to pH ≈ 7, and its volume was adjusted to 250 mL. This final solution was used as the titrant of a standard solution of potassium permanganate to quantify the nitrogen oxides absorbed.

Results and Discussion

Analytical Curves for Nitrite and Ammonium Determination by ATR–FTIR

To follow ammonium and nitrite consumption during NGS reactions, the mid-infrared spectra of these ions in water at different concentrations of both ions were acquired. A univariate calibration model that associates the area of the selected peaks with their respective concentrations was performed. Figure shows the wavenumber ranges in which water, ammonium, and nitrite presented their main absorption in the mid-infrared region.

Figure 1

Infrared absorption peaks in water solvent. The peak areas of ammonium and nitrite are highlighted.

To construct the analytical curves, the peak areas in the region of 1300–1100 cm–1 for nitrite and 1500–1410 cm–1 for ammonium were measured. Both absorption ranges were normalized in relationship to the water absorption maximum at 1641 cm–1. These curves were prepared from 17 solutions of each ion, as shown in Figure a,b, respectively. Both ions presented a very good correlation between the FTIR absorption peak area and the corresponding concentration.

Figure 2

Analytical curves for (a) ammonium and (b) nitrite solutions obtained at 5 °C by measuring the peak areas highlighted in Figure .

Real-Time Monitoring of the NGS Reaction with ATR–FTIR and Calorimetry

The good fit between the peak area and concentration expressed in the calibration curves allows real-time following of the consumption of ammonium and nitrite in the NGS reaction. If the reaction follows a stoichiometric ratio of 1:1 (eq ), the decrease of both concentrations should occur at the same rate.

However, because there is a possibility of nitrite consumption by a side reaction producing NO (eqs and 3), the observation of a higher rate for the nitrite consumption in relationship to ammonium may indicate this occurrence. Figure shows an example of an NGS reaction in which this behavior was observed.

Figure 3

Real-time ATR–FTIR monitoring of the NGS reaction for the following experimental conditions: temperature, 8.75 °C; [NH4+] and [NO2–] concentrations, 2.67 mol L–1; catalyst [HAcO] concentration, 0.13 mol L–1. (a) 3D plot of the ATR–FTIR real-time monitoring; (b) ammonium and nitrite concentration.

As expected, the water absorption was constant during the entire experiment, and the concentrations of ammonium and nitrite started decreasing as soon as the catalyst, acetic acid, was added. It is easy to see in the graph that the nitrite concentration decreased faster than the ammonium, indicating the formation of NO. The concentration of NO produced in the NGS reaction, that is, the amount of gas produced per volume of solution, was calculated using eq , in which the variation of the concentrations of both ions was obtained from the data presented in Figure b.

Figure shows the production of NO and the consumption of ammonium and nitrite from this experiment.

Figure 4

Production of NO and ammonium and nitrite consumption during the NGS reaction for the following conditions: temperature, 8.75 °C; [NH4+] and [NO2–] concentrations, 2.67 mol L–1; catalyst [HAcO] concentration, 0.13 mol L–1.

According to the data in Figure , the NO production reached 6.5% of the initial concentration of nitrite. This quantity of such a harmful product should not be neglected for applications of the NGS reaction in the topside and subsea environments, indicating the need for a deeper evaluation of the reaction conditions that will minimize NO production.

Aiming at validating the results presented in this work, the amount of NO produced in the NGS reactions was also determined by absorbing the formed gas in an ammonium hydroxide solution according to a methodology described in the literature.[15] The method is based on the conversion of nitrogen oxides to ammonium nitrite (eq ).

The amount of ammonium nitrite is proportional to the nitrogen oxides and can be quantitatively determined using the standard potassium permanganate titration procedure.[27] To check if there was a significant difference between the NO values obtained by titration and those obtained by infrared spectroscopy, a paired t-test was carried out in which the null hypothesis that there was no difference between the experimental means obtained from the different techniques was evaluated.

According to the results, there is an equivalence between the two techniques. Therefore, the hypothesis that there is no difference between the NO measurements for ATR–FTIR and titration is valid. From now on, the amount of NO produced will be accepted as equal to the difference in nitrite and ammonium consumption measured by ATR–FTIR. Section S1 of the Supporting Information presents the experimental conditions (Table S1), the NO values obtained by titration and ATR–FTIR, and statistical results of the t-test (Table S2).

By considering that the heat released in the reaction is proportional to the conversion of the reagents into products, the monitoring of the heat flow by a reaction calorimeter should also reflect the yield of the NGS reaction. Figure presents the comparison of the heat measurement with the consumption of the NGS reagents using FTIR. It should be noted that the evaluation of the reaction yield by using nitrite consumption showed a higher value than what was measured for ammonium or heat flow, and this difference corresponds to the NO formation.

Figure 5

Yield of the NGS reaction determined by heat flow calorimetry and ATR–FTIR. The experimental conditions of this experiment were temperature, 8.75 °C; [NH4+] and [NO2–] concentrations, 2.67 mol L–1; catalyst [HAcO] concentration, 0.13 mol L–1.

Preliminary Evaluation of the Reaction Conditions in the Heat Generation and Nitrogen Oxide Production

Catalyst Concentration

The reaction of ammonium and nitrite ions is acid-catalyzed, so the pH should exert an influence on the release of heat in NGS reactions. Figure shows the evaluation of the effect of acetic acid concentration in the reaction yield measured by calorimetry.

Figure 6

Effect of catalyst concentration in the (a) pH and (b) reaction kinetics. The experimental conditions of this experiment were temperature, 8.75 °C; [NH4+] and [NO2–] concentrations, 2.67 mol L–1.

Prior to the addition of the catalyst (where pH ≈ 6.60), the reaction does not occur. However, with its addition, the pH of the medium drops abruptly, and the reaction begins. As expected, the increase of catalyst concentration led to an increase in the kinetics of heat release that, in turn, is proportional to the reaction yield.

As was already emphasized, a drawback of the NGS reaction is NO production. Figure presents the NO formation for the same experiments presented in Figure . The increase in catalyst amount led to an increase in the production of heat (Figure b), but this greatest heat release caused an enormous increase in the production of NO (Figure ). Similar to what is observed for the reaction of ammonium and nitrite ions, NO generation is also dependent on the acidity of the medium, which influences the concentration of nitrous acid.

Figure 7

NO production as a function of catalyst concentration. Experimental conditions: temperature, 8.75 °C; [NH4+] and [NO2–] concentrations, 2.67 mol L–1.

By manipulating only the pH of the medium through adjustment of the acetic acid concentration, it is possible to obtain a variety of profiles of heat release and generation of oxides. The results shown in Figures b and 7 clearly indicate that the higher the catalyst concentration (lower pH values), the higher the rate of heat release and nitrogen oxide generation. Additions of lower amounts of catalyst, in turn, decrease NO generation, while the release of heat also drops significantly, somewhat undesirably.

An alternative to achieve good calorimetric yields with lower NO production is the addition of the catalyst over time, a situation illustrated in Figure . When the catalyst is divided into two aliquots, in which the second is added 2 h after the first, there is a greater control of the pH of the medium: the second aliquot is added at a point when the nitrite concentration is lower than the initial one, which limits NO generation.

Figure 8

Effect of the distribution of the catalyst over time in the (a) pH, (b) reaction kinetics, and (c) NO production. The experimental conditions of these experiments were temperature, 8.75 °C; [NH4+] and [NO2–] concentrations, 2.67 mol L–1; final catalyst [HAcO] concentration, 0.10 mol L–1. The orange experiment corresponds to the total addition of the catalyst, while the purple experiment corresponds to the addition of the same amount but in two different aliquots (the second half was added 2 h after the first).

Conversely, the literal addition of the catalyst sharply reduces the pH in a condition in which the nitrite concentration is maximal, which increases the production of nitrogen oxides. However, by adding the catalyst stepwise, the generation of heat is delayed.

Initial Concentrations of Ammonium and Nitrite

The ratio between ammonium and nitrite concentrations (R = [NH4+]/[NO2–]) was evaluated on three different levels. Figure presents the results.

Figure 9

(a) NGS yield curves based on ammonium consumption and (b) NO production. The experimental conditions of these experiments were temperature, 5 °C; catalyst [HAcO] concentration, 0.13 mol L–1.

A preliminary analysis of these curves indicates that the nitrite concentration has a greater influence on the heat generation than the ammonium concentration. Thus, the experimental condition in which the [NH4+]/[NO2–] ratio is 0.25 is the one in which the greatest and fastest heat release occurs. However, because of the larger excess of nitrite, the nitrogen oxide production reaches its highest percentage (Figure b).

On the other hand, because the catalyst concentration was the same for these experiments, it can be assumed that the lower generation of oxides in the case [NH4+]/[NO2–] = 1.00 is due to the smallest amount of nitrite in the medium and its faster consumption by the NGS, which limits the availability of nitrous acid to decompose (see eq ).

Design of Experiments (DoEs): Evaluating the Effect of Different Reaction Conditions on the Production of Heat and Nitrogen Oxides

Knowing that the catalyst amount and [NH4+]/[NO2–] ratio influence heat generation as well as production of nitrogen oxides, it is advisable to investigate experimental conditions that favor the production of nitrogen in detriment to the decomposition of nitrous acid.

Essentially, the following must be emphasized:

The decrease of the [NH4+]/[NO2–] ratio increases the rate of heat generation, but very small ratios limit heat production and significantly increase NO production.

Increasing the catalyst concentration increases the rate of heat generation but significantly increases NO production. In contrast, the stepwise addition of the catalyst can minimize NO production, although retarding the generation of heat.

Raising the temperature leads to speeding up of the reaction kinetics. Would that cause an increase in NO production?

Aiming to evaluate heat and NO simultaneously, a central composite experimental design[28] was performed, whose factors and their levels are presented in Table . Regarding the factors, the “partition” corresponds to the division of the total catalyst amount into two aliquots, where the second half is added P hours after the first (this time is defined by the value of this variable). This factor was included to evaluate the catalyst distribution over time, which avoids a sharp drop in pH under conditions of higher nitrite concentration.

Table 2

Factors, Their Levels, and the Responses of the DoE Performed

factors	–2	–1	0	+1	+2
R: [NH4+]/[NO2–]	0.50	0.75	1.00	1.25	1.50
C: [HAcO] (mol L–1)	0.01	0.04	0.07	0.10	0.13
T: temperature (°C)	1.25	5.00	8.75	12.50	16.25
P: partition (h)	0.0	1.0	2.0	3.0	4.0

The response y1 was measured at the end of 11 h of reaction, enough time to ensure that the reaction had already ceased, allowing the maximum potential of heat generation of the NGS (maximum yield) to be evaluated. It is also possible to evaluate the rate of heat generation through the response y2: high values indicate a large release of heat in short times.

The models were evaluated by multiple multivariate linear regression, in which a model was obtained for each of the three responses. The models initially contained first- and second-order terms, with interactions between variables and quadratic terms of each factor individually. ANOVA was performed to analyze the generated models, and the least contributing variables were eliminated by backward elimination with α = 0.05. Section S2 of the Supporting Information presents details of the results. The final models for each response are presented in eqs –11.

For all responses, the influence of temperature was complex; it depends on the catalyst concentration and the ammonium/nitrite ratio for the NO generation (terms +0.92RT and +0.31CT in eq ) and on the ammonium/nitrite ratio for the reaction yield (terms −1.92RT in eq and −0.93RT in eq ). As for the effect of the catalyst partition on the reaction yield, the absence of this variable in the final model of y1 was observed (eq ), while it appeared in the final model of y2 (eq ). This indicates that by adding the catalyst at different times, there is a delay in the release of heat, but the potential of heat generation is independent of this factor and is achieved in sufficiently long periods. The partitioning of the catalyst also seems to influence the NO generation, in which the addition of the catalyst at different times controls this side reaction.

A common condition of application of NGS is that in which the temperature is about 5 °C and the catalyst is added completely at the beginning of the reaction. This scenario simulates conditions found in ultradeep subsea applications, in which the NGS is used by the oil industries to fluidize low melting point solids formed in equipment related to subsea wells. For this scenario, Figure indicates the response surfaces obtained by the constructed models.

Figure 10

Response surfaces for (a) yield after 11 h, (b) yield after 1.5 h, and (c) NO generation for the case in which the temperature is 5 °C and the catalyst is added fully at the start of the reaction.

The reaction conditions that converge for a rapid and large heat release (e.g., R = ±2 and C = +2) also lead to a higher generation of nitrogen oxides. Thus, each response has a combination of factor levels that optimizes its value. Table lists these combinations.

Table 3

Combinations of Factor Levels That Optimize Each Response

effects	R	C	T	P	y1 (%)	y2 (%)	y3 (%)
maximizes y1	2	1	–1	–2	85.2	38.3	12.3
maximizes y2	2	2	–1	–2	81.9	40.6	12.7
minimizes y3	0	–2	–1	–2	16.5	–2.6	1.32

To maximize the overall reaction yield, using an excess of the reagents can lead to a yield of over 80% but with a NO generation of over 10%. It should be noted that even though the model’s global maximum is for R = +2 (excess of NH4+), using R = −2 (excess of NO2−) achieves basically the same results. This can be seen in Figure a by the symmetry of the surface curve over the R variable.

Minimizing NO generation, on the other hand, requires the minimum amount of the catalyst added to solution, with an equivalent concentration of ammonium and nitrite. This condition can decrease the amount of NO generated in around 90%, with a final NO yield of 1.32%. However, the overall NGS reaction also has a decrease in its yield, going from 84.6 to 16.5%. This means that decreasing the amount of generated NO also leads to less generated heat. Because the level of one factor can improve one response and worsen another, the optimization of the NGS process can be made by choosing levels that make heat generation much larger than NO generation. Thus, an optimization can be made to maximize a different function that takes into account, simultaneously, y1 and y3.

This objective function, f, can be written as f = y1/y3, and the optimization for fixed y1 values leads to the minimization of y3 for the maximum set y1. This can be thought of as the estimation of how much heat can be generated with the least amount of NO produced as a side reaction. It should be noted that the maximum possible NGS yield is presented in Table and is 85.2%, with a NO yield of 12.3%. In the same sense, the NO generated can be reduced to a minimum yield of 1.32%, leading to a maximum overall NGS yield of 16.5%. These two points are the limit of the function f, representing the set of maximum and minimum yields that can be achieved for the subsea conditions. Optimizing the function for different set values of y1 leads to Figure a.

Figure 11

(a) Optimization of the objective function f for different set values of y1 and (b) values of the variables R and C.

It can be seen that decreasing the amount of generated NO leads to a decrease in the NGS yield, which indicates that less heat should be used in the application. However, it can be noted that there are two regions with different slopes, indicating that different phenomena control the system’s behavior. Evaluating how the R and C variables change for this function leads to Figure b.

Initially, going from the maximum to the minimum NGS yield in the gray region of Figure b, it can be seen that to decrease the NO generation, the only variable that needs to be changed is the reagent ratio because only this factor changes, while the factor [HAcO] remains approximately constant (∼0.09 mol·L–1). This means that changing the catalyst concentration does not increase the y1/y3 ratio, which means losing heat at a cost of not reducing most of the NO generated. This first region can be considered a “reagent ratio controlled” and goes up to an overall yield of around 50%. At this point, the reagent ratio is 1, indicating that ammonium and nitrite concentrations are the same.

To decrease even further the NO generation, the catalyst concentration must be reduced up to its smallest value of 0.01 mol·L–1. This region can hence be considered “catalyst controlled”. However, reducing NO until this limit is not the ideal scenario because reducing it leads to a bigger loss in the generated heat, which means that more solution must be added to the system to achieve the same amount of heat. Increasing the amount of solution increases the amount of NO that can be generated, which is not ideal.

Because the minimum amount of NO is in the catalyst-controlled region, a plot of the function f against C shows how much concentration needs to be decreased to maximize f. Results are presented in Section S3 of the Supporting Information. It is seen that decreasing from C = 0 to C = −2 does not increase much the function f, which does not bring much gain to avoid the NO problem. Hence, using a set of R = 0 and C = 0 leads to the highest y1/y3 ratio, in which the most NO is reduced in relationship to the reduction in the amount of heat generated.

For this configuration, the overall NGS yield is 40.9%, and the NO yield is 4.14%. For this combination, by changing the partition value to P = 2, the NO generation drops by about 47%, while the heat generation is delayed (y2 drops about 20%). As discussed, the yield at the end of 11 h does not change once the overall reaction yield is not changed by the partitioning of the catalyst.

Conclusions

Minimizing NO production in the NGS requires careful control of ammonium/nitrite ratio and catalyst concentration in order to ensure the lowest environmental impact in the application of this technique, without prejudice to heat generation. Because these factors exert opposite influences on heat and NO generation, a balance between them is essential to achieve the greatest and fastest heat release with the lowest NO production. These conditions correspond to a ratio of ammonium/nitrite equal to 1 and a catalyst concentration of 0.07 mol·L–1, for a case in which the temperature is 5 °C (seafloor level). Still, the addition of the catalyst over time is a good strategy to decrease NO generation without affecting the maximum potential of heat generation of the NGS.