ANALYST SEPTEMBER 1986 VOL. 111 Use of Solid Boric Acid as an Ammonia Determination of Nitrogen Darryl 0. Siemer WINCO PO Box 4000 ILF 0-25 Idaho Falls ID 83403 USA 1013 Absorbent in the An inexpensive trap - de-mister assembly utilising solid crystalline boric acid as an ammonia absorbent was developed to replace the specialised trap - condenser - impinger apparatus normally used in Kjeldahl-type distillations. The boric acid is subsequently dissolved in water and a final determination of the ammonium ion is made either by acidimetric titration a conductance measurement or spectrophotometric measurement after the addition of Nessler‘s reagent. The apparatus fits into the neck of a calibrated flask during the distillation and is itself constructed from disposable plastic pipette tips glass-wool and rubber tubing.Both the high surface area and the tortuosity of the paths through the randomly packed bed of crystals makes solid boric acid an efficient ammonia absorbent. The system is useful for small-scale separations only because, when the volume of solution from which the ammonia is distilled exceeds 25 ml the acid crystals become saturated with water and begin to dissolve before all of the ammonia is trapped. Keywords Solid-state ammonia trap; boric acid; Kjeldahl distillation; nitrogen determination Taras’s review of nitrogen determination procedures empha-sises the importance of what will hereafter be referred to as the “Kjeldahl approach.”l The two main steps in the Kjeldahl approach are firstly a chemical conversion of the nitrogen in the sample to ammonia or ammonium ion and secondly the distillation of the ammonia into an acid liquid trapping agent.The ammonia in the trapping agent is then determined by any of a number of suitable finish techniques. Whereas the original Kjeldahl technique was developed for “organic nitrogen” determinations the same basic approach can be used for determining practically all of the chemical forms of nitrogen if the distillation step is preceded by a suitable chemical pre-treatment of the sample. The bottleneck in many of these analytical procedures is the distillation - separation process not the initial chemical conversion step. Traditionally this distillation step is per-formed with specialised all-glass stills that have a water-cooled condenser the free end of which is immersed in the trapping solution (e.g.the Pregl - Parnas - Wagner apparatus). These distillation systems typically possess high internal surface areas and considerable dead volumes. Consequently a relatively large volume of water must be boiled over to flush the ammonia released from the sample digest quantitatively through to the trapping agent. This paper discusses a simple and inexpensive solid-state ammonia trap which significantly reduces the time necessary for these separations in many situations. An outline of some practical applications is included in order to illustrate the utility of the system. Experimental Apparatus Fig. 1 shows an ammonia trap assembly inserted into the neck of the distillation flask. It consists of a standard 1-ml disposable plastic micropipette tip filled approximately one third with boric acid crystals (about 0.5 g).Several batches of boric acid from different manufacturers were used with essentially equivalent results. Crystal sizes were such that more than 95% of the boric acid was retained between sieves passing 850 and 150 pm particles (Tyler 20-100-mesh screens). A plug of glass-wool supports the crystals and a plastic “chimney” prevents steam from pushing the plug of moist crystals out of the top of the trap. The chimney is made by cutting both the top and bottom off another pipette tip. A plastic retainer cut from yet another pipette tip is forced into Fig. 1. Trap - de-mister assembly. A Chimney; B pipette tip used for body of trap; C boric acid crystals; D retainer; E glass-wool; F, tip used for de-mister; G gum - rubber gasket; H neck of distillation flask the trap body over the glass-wool before the boric acid is added.This prevents the glass-wool from being washed out along with the boric acid crystals after the distillation step has been completed. The ammonia trap is inserted into another pipette tip loosely filled with glass-wool that serves as a “de-mister” to intercept the particles of spray released during the boiling of the sample digestate. The de-mister serves the same purpose as the trap usually incorporated in commercial Kjeldahl systems but it provides the necessary surface area with far less dead volume and solution “hang up.” A short length of gum-rubber tubing is forced over the bottom of the de-mister to serve as a gas-tight gasket between it and the neck of the flask.Both the de-mister and the trap assembly can be re-used an indefinite number of times. After the boric acid used in a prio 1014 ANALYST SEPTEMBER 1986 VOL. 111 determination has been removed the system is prepared for the next by connecting it to a sink-aspirator vacuum source and then pulling water acetone and finally air through it. Distillation Procedure Standard 25- or 50-ml calibrated flasks and a magnetic stirrer -hot-plate were used for both the sample preparation and distillation steps. To perform an ammonia distillation the contents of the digestion flask are first cooled if necessary, and then diluted to approximately 5 ml with water.If a magnetic stir-bar was not used during the sample preparation step one is now added to the flask. Alternatively a few boiling chips can be added. This is necessary because “bumping” can cause expulsive loss of the trap - de-mister assembly and/or its contents. Enough strong base to provide an excess of at least 2-3 mmol of free hydroxyl ion is then added and the filter - de-mister assembly is immediately inserted into the neck of the flask. The flask is placed on to a pre-heated stirrer - hot-plate and its contents are rapidly brought to boiling-point. The trap is removed after an approximately 1.5 min distillation time. The chimney is pulled out of the trap and any boric acid crystals adhering to it are rinsed into a small beaker containing hot water.The contents of the trap are then back-washed into the same beaker using a plastic wash-bottle with a spout modified to fit snugly over the bottom tip of the trap. Finish Techniques For sample digestates containing relatively large amounts of ammonium ion a titration of the dissolved boric solution proved satisfactory. Standardised hydrochloric acid was added with a 2-ml micrometer burette (Cole Parmer) until the methyl orange visual end-point (ca. pH 4) was reached. For low concentrations a spectrophotometric finish tech-nique was investigated. This involves the neutralisation of the boric acid with a slight excess of strong base (about 1.1 equivalents of NaOH per mole of boric acid in the trap) followed by the addition of Nessler’s reagent and dilution to a suitable volume.The absorbance of the fully developed yellow chromogen was then determined in a spectrophotometer at 425 nm in a l-crn path-length cuvette. An alternative and usually superior finish technique consists of the measurement of the conductivity of the boric acid solution. Boric acid is an extremely weak acid (pK = 10.1) and consequently its aqueous solutions are very poor electrical conductors. On the other hand ammonium borate is a typical ionic salt and gives rise to a relatively large electrical response. For this work the detection module of a Dionex Model 2000i ion chromatograph was re-plumbed so that the columns were by-passed and the solution could be sucked directly through the conductivity cell with a plastic syringe. After the cell is filled with the solution the measurement is made under “zero-flow” conditions.Results and Discussion The initial testing of the solid-trapping agent concept was performed by simply drawing headspace fumes from a reagent “ammonia” bottle through one of the traps with a syringe and then blowing the contents of the syringe past the investigator’s nose. The fact that no ammoniacal odour could be detected indicated that boric acid is an effective ammonia absorber. Table 1 outlines a study performed to determine the effect on ammonia recoveries of using varying distillation times. Aliquots of a standard ammonium sulphate solution were used so that no prior sample preparation step was needed. The amount of ammonium ion taken was chosen to be relatively large (0.1392 mmol) so that the precise titrimetric finish could be employed.The “primary standard” on which the recovery Table 1. Ammonia recovery as a function of distillation time. In all instances the sample consisted of 0.100 ml of 1.392 M ammonium sulphate solution and a total solution volume of approximately 5 ml containing 10 mequiv. of free hydroxyl ion was boiled Mass of 20 0.35 85.0 30 0.70 99.85 45 0.82 100.01 90 1.40 99.36 150 2.03 100.20 240 3.30 99.71 * The interval from when the condensate first appeared in the distillate/g Recovery YO Time/s* de-mister to when the filter was removed. values in both this and the following tables are based is the sodium carbonate used to standardise the hydrochloric acid titrant. In this series of experiments the total solution volume in the distillation flask prior to placing it on to the hot-plate was approximately 5 ml.The flask was weighed both before and after the distillation to determine the amount of solution boiled away and passing through the de-mister - trap assembly. The total time required for each distillation was approximately 30 s greater than the figure listed in the first column because that interval was required to bring the solutions to the boil. The data indicate that when solution volumes are relatively small essentially complete ammonia recovery is obtained within 30 s of the onset of rapid boiling. In this instance this corresponds to 10-15% of the solution in the flask. As might be expected a series of further experiments performed with different volumes and compositions of the solutions in the distillation flask revealed that the distillation times required for quantitative recoveries were a function of both the volume of the solution being boiled and of its total salt concentration.As a rule of thumb quantitative ammonia recovery is achieved by the time that 15% of the aqueous phase has been boiled away. Higher total salt contents raise the boiling-point of the solution which in turn enhances the volatilisation of the ammonia relative to that of the water. The upper limit to the solution volume from which the ammonia can be effectively trapped is determined by the solubility of the boric acid in the water absorbed/condensed in the trap during the distillation process.After complete water saturation of the boric acid occurs a portion of the resulting boric acid solution formed subsequently might either drip back into the de-mister or be carried out of the top of the chimney as a spray. This effectively limits the amount of steam that can be passed through the trap to no more than that generated by boiling away approximately 3-4 ml of the solution in the flask. With high salt concentrations (the equivalent of approximately 4 M NaOH) present ammonia can be quantitatively recovered from a maximum digestate volume of approximately 25 ml. However to provide a good margin of safety in this respect solution volumes of no more than 15 ml are recommended. Tables 2 and 3 give the results of some applications of these traps to nitrogen determinations in a number of different types of compounds.In the experiments described in Table 2, enough of the compound was weighed into a 25-ml calibrated flask to provide 0.3-0.5 mmol of “amino - amido” nitrogen. A 0.5-ml volume of 9 M sulphuric acid was added and the sample was then digested for 5 min on a pre-heated hot-plate with a surface temperature of approximately 400 “C. Under these conditions the water immediately boils off and the sample digests in quietly refluxing concentrated sulphuric acid at a temperature of approximately 300 “C. No digestion catalyst was used or needed with these readily hydrolysed compounds ANALYST SEPTEMBER 1986 VOL. 111 1015 Table 2. Recoveries of nitrogen from various compounds. Samples were digested with 0.25 ml of sulphuric acid for 5 min Compound Recovery Yo Iron(I1) ammonium sulphate .. 100.85 Thiourea . . . . . . . . . . 98.87 Ammonium sulphate . . . . . . 99.97 Sulphamic acid . . . . . . . . 99.97 Table 3. Recoveries of nitrate-nitrogen from various compounds. Samples containing approximately 0.4 mequiv. of nitrogen were reduced with aluminium powder in strong base prior to the distillation step. The relative standard deviation of the nitrate determinations performed at this concentration level was approximately 1 YO Compound Recovery YO Nitric acid . . . . . . . . . . 99.89 Cadmium nitrate . . . . . . 99.85 Calcium nitrate . . . . . . . . 99.67 Sodium nitrate . . . . . . . . 100.10 The distillation - separation procedure outlined above was then applied.The excellent nitrogen recoveries obtained for the hydro-lysed sulphamic acid samples were of special interest to this laboratory because the original focus of this project was to develop a rapid procedure to determine sulphamate that did not rely on the often troublesome titrimetric determination of nitrite. Indeed in many instances the accuracy and precision of sulphamic acid determinations performed with the “Kjel-dahl” approach proved to be much better than could be routinely obtained by a more conventional redox titrimetric method involving nitrite.2 However as a general-purpose analytical method for sulphamate at this facility the technique proved to have a major weakness. This is that any concomitant nitrate severely interferes during the hydrolysis step.Nitrate (or one of its decomposition products in the hot sulphuric acid) apparently oxidises some or all of the sulphamate’s “amido nitrogen” to elemental nitrogen. Table 3 shows the results of nitrate determinations per-formed using the same final separation and finish steps after prior reduction of the sample with metallic aluminium.3 The procedure involved placing a suitable amount of sample (containing 0.3-0.5 mmol of nitrogen) into a 50-ml flask and then dissolving it in about 2-3 ml of water. Approximately 150 mg of powdered aluminium followed by 2 ml of 10 M NaOH were added and the ammonia trap - de-mister assembly was immediately seated into the neck of the flask. After the initial vigorous reaction had subsided the flask was placed on the hot-plate and the ammonia remaining in the flask distilled as described above.Recoveries were again excellent, The spectrophotometric finish approach can be recommen-ded only for those situations in which its excellent analytical sensitivity is actually required by the problem at hand. Nessler’s reagent is notoriously difficult to use in high-salt sample solutions and in this work also tended to give erratic results unless considerable care was taken to reproduce exactly all the relevant conditions in preparing the solutions for the spectrophotometer. The absolute and relative amounts of boric acid and base used solution temperatures reagent addition technique reagent volume etc. all need to be reproduced carefully for good over-all analytical precision.Preliminary experiments also revealed that the “phenate” procedure often recommended for water analyses was even less satisfactory.4 Both boric acid and borates apparently inhibit the development of the indophenol chromogen. The most convenient finish technique proved to be the conductance measurement. It was developed to serve as a substitute for the more troublesome Nesslerisation technique for low-nitrogen samples. The mean conductance of a series of six blanks run through the distillation procedure was 7.15 pS and the standard deviation was 0.53 CIS. In each instance 0.50 k 0.02 g of boric acid was used in the traps and the final solution volume was 50 ml. Essentially identical conductance values (a mean of 7.21 pS with a standard deviation of 0.48 pS) were obtained with 1% boric acid - water solutions prepared directly without the intermediate distillation and transfer steps.This indicates both that the de-mister does a very effective job of preventing spray carry-over and that, with reasonable care avoidance of contamination of the boric acid with miscellaneous salts throughout the handling process is possible. The analytical response generated by ammonia added to these solutions is approximately 5.6 pS per p.p.m. of ammonia and the response is linear to at least the 1000 pS maximum conductivity range of the detector used for this work. This translates to an analytical range from approximately 0.25 p.p.m. (the detection limit based on the 30 criterion) to about 180 p.p.m. a range encompassing trace to macro levels of nitrogen in the original sample.Conclusions The limited solution volume from which ammonia can be quantitatively recovered probably constitutes the most serious limitation of the ammonia traps. This factor effectively limits their application to relatively small-scale work only. However, for many applications the fact that their use permits good analytical results to be obtained quickly with otherwise non-specialised glassware and equipment makes them an attractive alternative to classical Kjeldahl still set-ups. The most serious limitation of the conductance finish technique is its lack of specificity. Any ionic species present either in the original boric acid or inadvertantly picked up during transfer steps can give serious blank problems. However with reasonable care it is not too difficult to avoid contamination. References Taras M. J. in Boltz D. Editor “Colorimetric Determi-nation of Nonmetals,” Interscience New York 1958 Chapter Whitman C. L. Anal. Chem. 1957 29 1684. Bartow E. and Rogers J. S . Waf. Sew. Ser Ill. Univ. 1909, 7 14. American Public Health Association American Water Works Association and Water Pollution Control Federation “Stan-dard Methods for the Examination of Water and Wastewater,” Fourteenth Edition American Public Health Association, Washington DC 1975 pp. 416-417. IV pp. 87-91. Paper A6143 Received February 12th 1986 Accepted April 2nd 198