Determination of Nitrogen Trichloride in Water
---------Arthur T.Palin---------
A contribution submitted to the JOURNAL on Mar. 14, 1968, b::,'Arthur
T. Palin, Chief Chemist, jli euicastle & Gates/wad TVater Co., N eto-
castle, England.
F the chloramines that may form In practice the need to determine
O during the chlorination of water,
it is generally accepted that nitrogen
nitrogen trichloride rarely arises, but
its behavior seemed of sufficient in-
trichloride is the most difficult to de- terest to warrant further study. The
termine specifically. The results ob- results now obtained suggest that some
tained for that compound appear to be of the discrepancies reported previ-
too low when compared with corre- ously could have been due to the ef-
sponding standard iodometric results. fect of aging on the stock solutions.
FAS Method Experimental
In the FAS method using diethyl- Stock solutions of NCl3 were pre-
para-phenylene diamine indicator pared according to the procedure given
(DPD),1 the only response from NCI3 by Chapin." Appropriate amounts
in the normal stepwise titration lies in were added subsequently to distilled
the dichloramine fraction. Its appear- water to give total available chlorine
ance here rather than with free chlo- concentrations of up to 5 mg/l. These
rine is advantageous because those dilutions were then examined by the
compounds likely to prove objection- F AS-DPD procedure with the stan-
able in practice are thus grouped to- dard ioclometric method used in some
gether. of the experiments as a check on total
In the development of the FAS-DPD available chlorine. Monochloramine
procedure, however, it was demon- being usually present in trace amounts
strated that NCI3 could be caused to was included with free-chlorine.
respond in the free-chlorine fraction if The amount of KI used to activate
first activated by the addition of a NCI3 to cause part of it to appear with
small amount of potassium iodide. free-chlorine was the same as that nor-
The evidence indicated that the amount mally used to activate monochlorarnine
obtained corresponded to 50 per cent
in the straightforward F AS titration.
of the total available chlorine of the
To remove N'Cl; when required, an
NCls. Most of the remainder con-
tinued to appear with dichloramine, extraction procedure was followed
although in only about 40 per cent, using carbon tetrachloride. Loss of
instead of the expected 50 per cent, dichloramine was allowed for by ap-
of its available chlorine. plying a correction factor based upon
847
848 A. T. PALIN Joltr.AWWA
the partition coefficient for this com- was almost entirely in molecular form
pound given by Chapin. and thus was extractable by CCI •.
In the light of the foregoing notes Readings under 3A were therefore of
the experimental procedures are now only trace amounts and are omitted
presented in Table 1. from the foIlowing tabulations. They
It was considered possible from these do not in any case enter into the calcu-
determinations to evaluate the propor- lations relating to no, (Table 2).
tions :r and y of the total NCIs content
appearing in Procedure 2, Fraction A Older Stock Solutions
and in Procedure 1, Fraction C. It will be observed in experiment 2
Thus that with aging of the stock NCI3
(x) NCb:= 2A - lA solution, the proportion appearing with
and dichloramine becomes progressively
(y) NCb:= l C - 3C less (Table 3).
As there was already evidence indi-
cating that :r corresponded to SO per Standing NCl3 Solutions
cent, the most satisfying results would In Experiment 3, instead of aIlowing
have been where y also equalled 50 the stock NCI3 solution to stand and
per cent of the total available chlorine preparing dilutions from it at intervals,
content of NCIs, so that all could be the initial dilution itself was all owed to
accounted for. Although such result stand (Table 4).
was often obtained, there were many With the diluted NCI3 solution the
puzzling discrepancies in which y ap- effect observed 111 Experiment 2 is
peared to be very much less than 50 not as marked.
per cent. Eventually it became clear
that the nature of the result depended Distilled Water Solutions
upon the age of the stock NCI3 solution.
A solution of NCl3 in distilled water
It would seem that slow secondary re-
was prepared by adding 10 mg/l CI
actions occur in Chapin's solution to
and 0.5 mg/l N (as NH.CI) and al-
produce amounts of some other un-
lowing to stand 30 min in Experi-
identified chlorine-nitrogen compound
ment 4.
or complex with rather different prop-
The free-chlorine readings after CCI.
erties from those of N C13•
extraction were now appreciable and
Fresh Stock Solutions are included in the tabulation in Table
5.
Because of the low pH of the solu- Taking the 3-day results, the total
tions in Experiment 1 the free chlorine available chlorine may be calculated as
follows: free-chlorine (IA) =2.54; di-
TABLE 1 chloramine (3C) = 0.10; nitrogen tri-
Fraction of DPD Method chloride (2A - lA) + (lC - 3C) =
1.59; the total = 4.23.
Procedure A c The corresponding figure by the
1. Normal CI, NHC)' plus (1') NCb iodometric method was 4.32 mg/l.
2. Kl added C)' plus (x) NCb ..~-
3. CCI. some Cl a NHC'; (corrected) A second solution prepared with 15
extracted
mg/I CI and 1 mg/I N gave the follow-
lui. 1968 NITROGEN TRICHLORIDE DETERMINATION 849
TABLE 2
Fresh Stock Solutions of NCl3
Readings for Various Dilutions-mg/l Ct
IA IC 2:\ 3C 2A-IA IC-3C
0.18 0.54 0.37 0.35 0.19 0.19
0.18 0.53 0.46 0.24 0.28 0.29
0.24 0.86 0.66 0.46 0.42 0.40
0.24 0.82 0.69 0.39 0.45 0.43
0.49 1.44 1.31 0.66 0.82 0.78
0.39 1.64 1.24 0.85 0.85 0.79
0.48 1.64 1.46 0.64 0.98 1.00
0.53 2.32 1.60 1.23 1.07 1.09
0.60 2.38 1.70 1.24 1.10 1.14
0.62 2.94 2.06 1.46 1.44 1.48
-- --
Mean 0.76 0.76
TABLE 3
Older Stock Solutions of NCl3
Readings for Various Dilutions-mg/l Cl
Age of Stock Solution
I
IA IC 2A .K 2A-IA I IC-3C
Several hours 0.15 0.59 0.49 0.36 0.34 0.23
Several hours 0.31 0.90 0.81 0.57 0.50 0.33
Several hours 0.21 0.86 0.72 0.48 0.51 0.38
1 day 0.13 0.53 0.42 0.36 0.29 0.14
1 day 0.23 1.14 0.96 0.80 0.73 0.34
1 day 0.33 1.31 1.09 0.87 0.76 0.44
2 days 0.19 0.83 0.87 0.71 0.68 0.12
TABLE 4
Standing Diluted NCl3 Solutions
Readings for Various Distr-ibut ions-e-mg/I Cl
Time of Standing
IA IC 2A 3C 2:\-1:\ IC-3C
Start 0.58 1.58 1.44 0.76 0.86 0.82
Several hours 0.48 1.61 1.30 0.75 0.82 0.86
1 day 0.41 1.45 1.20 0.79 0.79 0.66
2 days 0.35 1.31 1.12 0.78 0.77 0.53
850 A. T. PALIN Jour.AWWA
ing results after standing 10 days: free- present in trace amounts only, so that
chlorine, 0.79; dichloramine, 0.26; and Reading B (free chlorine plus 1110no-
nitrogen trichloride, 3.32. chloramine) was not required. A sec-
In this case the values of (2A - lA) ond determination of the first fraction
and (1 C - 3C) were 1.64 and 1.68. was made after preliminary KI treat"
It seemed clear that the observed ment. Reading D thus obtained cor-
discrepancies in earlier results were responded to (free chlorine plus i
associated with the aging of stock NCl3 N ClJ . Calcula tions were then car-
solutions prepared at high concentra- ried out as follows in Experiment 5:
tions. With available chlorine concen- free-chlorine = A; dichloramine = C -
trations more closely approaching those D; and nitrogen trichloride 2 (D - =
encountered in water chlorination, the A).
TABLE 5
NCt, in Distilled Water Prepared Direct From Low Concentrations of Cl and N
Readings for Various Distributions-mg/l Cl
Time of
Standing
IA IC 2A 3A 3C 2A-IA IC-3C
Start 3.58 2.76 5.63 3.06 0.68 2.05 2.08
1 day 1.80 1.60 3.30 1.85 0.12 1.50 1.48
3 days 2.54 0.90 3.33 2.3 0.10 0.79 0.80
indications were that the proportion of The iodornetric method was the stan-
nitrogen trichloride appearing in the dard acid-starch-iodide titration using
free-chlorine fraction after preliminary thiosulfate solution (Table 6).
KI treatment was equal to the propor-
tion appearing in the dichloramine frac-
tion of the normal procedure. Further-
more, it seemed reasonable to conclude TABLE 6
that together these amounts would ac- DPD and Iodometric Results for
count for the total nitrogen tricholride Dilutions of N Cl«
present. To check this point, a fur-
Iodo-
ther series of determinations was car- FAS·DPD Readings mgll CI metric
ried out, on which the calculated total
available chlorine contents by F AS- Cl NRC\' NCh Total Total
DPD were compared against the iodo-
0.10 0.22 0.92 1.24 1.15
metric results. 0.15 0.27 0.98 1.40 1.39
0.13 0.12 1.40 1.65 1.49
Comparison of Results 0.33 0.23 1.12 1.68 1.78
0.40 0.90 1.62 2.92 2.88
Determinations by the FAS-DPD 0.64 1.02 1.80 3.46 3.66
method consisted of Reading A (free- 0.68 1.09 2.14 3.91 3.75
chlorine) and Reading C (free chlo- 0.34 1.06 2.58 3.98 3.86
-- --
rine plus dichloramine plus i NCI3). Mean 2.53 2.50
Monochloramine was usually absent or
Jul. 1968 NITROGEN TRICHLORIDE DETERMINATION 851
Conclusions These conclusions apply only to the
low concentrations of NC1s such as
In comparing DPD results for chlo-
may be formed in the chlorination of
rine and chloramines in 'water with to-
waters or obtained by diluting freshly
tal chlorine figures obtained iodometri- prepared stock solutions. Solutions of
cally, a discrepancy arises whenever NCla that have been allowed to stand
NC1a is present, because only half of before dilution give inconsistent results.
this compound is recovered. This half
appears in the dichloramine fraction. References
The remaining half of the NC13, how- J. PALIN, A. T. The Determination of Free
ever, can be made to react in the free and Combined Chlorine in Water by
chlorine fraction by adding a small the Use of Diethyl-p-phenylene Diam-
ine. Jour. AWW A, 49 :873 (J ul.
amount of potassium iodide. In this 1957) .
way the total NC1a content may be 2. CHAPIN, R. 1\1. The Influence of pH
obtained and the dichloramine reading Upon the Formation and Decomposi-
tion of the Chloro Derivatives of Am-
corrected for the NC1a that appears monia. J. Am. Chem. Soc., 53 :912
with it. (1931).
