A SPECTROPHOTOMETRIC METHOD FOR THE DETERMINATION

OF AMMONIA IN WATER

by

JUI Wen-chun ( 瞿 文 俊 ）

A thesis submitted in partial fulfilment of the

requirement for the degree of

Master of Philosophy

in

The Chinese University of Hong Kong

1975

Thesis Committee:

Dr. P. K. Hon, Chairman

Dr. H. M. Chang

Dr. T. C. W. Mak

Prof. L. Ma

Prof. T. S. Ma, External Examiner

ACKNOWLEDGEMENT

The author wishes to express his sincere thanks to Dr. P. K. Hon for his

advice, guidance and encouragement during the course of his research and the

preparation of this thesis.

Special thanks are due to Prof. T. S. Ma and Dr. T. C. W. Mak for their

excellent advices.

Thanks are also due to all his friends and colleagues, especially

Mr. P. H. Fong and Mr. K. M. Sun, for helping to make his two years of

graduate study both stimulating and enjoyable.

He is also indebted to Mr. Y. Woo and Mr. T. Woo for making the

diffusion flasks.

Department of Chemistry JUI Wen-chun

The Chinese University of Hong Kong

June, 1975.

ABSTRACT

The Berthelot reaction method for the determination of ammonia in

solutions has been extensively tested. It is found that this method is

superior to the conventional Nessler' s method in many ways. Its coloured

product is truely soluble iii water and has an absorption maximum at 634 mono

Its experimental conditions are not as critical as in Nessler's reaction.

Under the optirnumw conditions found in present work, the sensitivity is 10

times higher than Nessler's method., and the product has nu ch higher stability.

There are still some inorganic and organic compounds interfering with

the eacticn. A simple extraction method based on diffusion is propesad.

An apparatus for this purpose is designed. Ammonia in sample solutions is

evolved by the addition of alkali. It is ao6orbed in an acidicsolution,

where the ammonia concentration can be determined by Berthelot reaction

method.

Compared to other extraction methods, simple diffusion method is much

simpler. No special technique and apparatus are required. The detection

limit is low (down to 0.). pprr ) and the results are quite accurate (with

root mean square error+ 0.009 ppm in the range 0.1--1 ppm). All inorganic

ions do not interfere with the method. Although some organic compounds

interfere, most of them are rare in natural water, and they show no inter-

ference when their concentration is under 1 ppm.

TABLE OF CONTEIMTS

PAGE

ACKNOWLEDGMENTii

ABSTRACT iii

LIST OF TABLES v

LIST OF FIGURESvi

I. INTRODUCTION 1

II. HISTORICAL REBIEW OF BERTHELOT REACTION METHOD 5

III.MECHANISM OF BERTHELOT REACTION 8

1. Introduction 8

2. Catalyst effect 9

3.Mechanisn10

IV. EXPERIMENTAL 19

1.Nessler's method 19

2.Direct Berthelot reaction method 21

3. Simple diffusion method 24

4. Apprication 27

V.RESULTS AND FISCUSSION 29

1. Nesslerts method 29

2. Direct Berthelot reaction method 61

3. Simple diffusion method 52

VI.APPLICATIONS61

1.Direct Nessler's method 61

2. Direct Bertblot reaction method 61

3.Simple diffusion method 62

4. Conclusion62

RFFERENCES 64

5

LIST OF TABLES

TABLE PAGE

1. The catalytic effect of some compounds on

Berthelot reaction. 35

2. The effect of complexing agent on the colouration

of Berthelot reaction. 36

3. Effect of the sequence of adding phenol and hypochlorite

in Berthelot reaction. 37

4. Effect of the time interval between hypochlorite

and phenol in the absorbance. 38

5. Interference study of inorganic ions on direct

Berthelot reaction method. 47

6. Interference study of organic compounds on direct

Berthelot reaction method. 49

7. Effect of the concentration of absorbent (NaHSO4). 53

8. The ammonia content in waste water of C.U.H.K. before

and after sewage treatment. 62

6

LIST OF FIGURES

FIGURE

1. Curves of the Bert'helot reaction using

nitroprusside and pre-alkalized nitroprusside

as catalyst.

2. Effect of pH value in the first step of Berthelot

reaction on the absorhance yield.

3.

4.

Reaction rate of monochloramine and phenol

Reaction rate of p-benzoquinonechloramine and

phenol at different pH.

6.

5. Absorption spectrum of indophenol solution.

The block diagram of the Hitachi model 323

UV-VIS-NIR recording spectrophotometer.

7. Apparatus of simple distillation method for

ammonia determination.

8. Calibration curve for Nessler's method.

18

(1 to 10 ppm ammonia concentration)

9. Calibration curve for Nessler's method.

(10 to 20 ppm ammonia concentration)

10. Calibration curve for Nessler's method.

(0.1 to 1.0 ppm ammonia concentration)

11.

12.

Absorption spectra of Nessler's reaction product.

Effect of phenol concentration on Berthelot

reaction.

13. Effect of the amount of sodium hypochlorite on

Berthelot reaction.

PAGE

11

13

14

17

20

26

30

31

32

33

39

39

7

14. Effect of the amount of sodium nitroprusside

on Berthelot reaction.41

15. Effect of pH on Berthelot reaction.41

16. Rate of colour development at different pH values of

Berthelot reaction. 42

17. Calibration curve for direct Berthelot reaction

method.

(0.1 to 1.0 ppm ammonia)44

18. Calibration curve for direct Berthelot reaction

method.

(0.01 to 0.1 ppm ammonia) 45

19. Effect of sample volume on ammonia absorption. 55

20. Effect of amount of 5 N NaOH on absorbance. 56

21. Galibration curve for simple diffusion metod.

(0.1 to 1 ppm ammonia) 59

1

I. INTRODUCTION

Ammonia is one of the most common compounds in nature linen animals

and plants die bacteria convert all the nitrogen in their bodies into

ammonia(1) That is the main source of ammonia in nature. Also there is

(2)a small amount of ammonia in human body

When nitrogenous materials of various types are digested in chemical

analysis (e.g. the Kgeldahl Method) the ultimate form obtained is always

ammonia. Therefore the determination of ammonia is the most important method

in chemical analysis for nitrogen in organic compounds.

To determine large amount of ammonia in solution direct titration with

a standard acid solution using methyl red as indicator is usually employed

(3)provided that no other basic constituents are present. Ammonia can also

be determined gravimetric ally by precipitation with chloroplatinic acid

(3)

to form ammonium hexachloroplatinate (NR4)2PtCl6.

For very dilute ammonia solutions the conventional method of deter-

(4)mination is Nesslers method which was developed by Nessler in 186

When the Nessler reagent (an alkaline solution of mercuric iodide) is added

to a solution containing ammonia either in combined or in free forms brown

coloured NH Hg I is produced,which can be determined spectrophotometrically2 2 3

(5)at the wave-length hlO nm. For higher ammonia concentration (up to 20 ppm)

longer wave-length can be used but in sacrificing of some accuracy. The

preparation and condition of Nessler reagent should be carefully controlled

in order to get reliable results. During Nesslerization numerous factors

such as pH temperature salt concentration rate of mixing and time for

(6,7)standing etc. , all can cause erratic results. Nessler's method is also

subjected to serious interference by.many kinds of chemicals j such as(8)

2

the appearance of yellow or green colour which is caused by a number of

aliphatic and aromatic amines and the onset of turbidity which is caused

by aldehydes acetone alcohols organic chloramines and some other unknown

organic compounds. Ions insoluble in alkaline solution or producing

precipitates with iodine and mercury also cause turbidityc.

Among all. other proposed methods for micro determination of ammonia,

the ninhydrin method(9) is not specific to ammonia. All amd.no acids can

give a similar colour reaction with the reagent(10) hence it is seldom

used in the quantitative analysis of ammonia.The pyridine-pyrazolone

(11)method proposed by Kruse and Mellon is subject to interference from

silver,copper,sine,and iron in excess of 50 ppm. Cyanide thiocyanate

and cyanate also form coloured compounds with the reagents and must be

removed. Moreover the reagent is unstable the absorption does not follow

Beer's law and the coloured product should be extracted with carbon tetra¬

chloride for the measurement. Although the sensitivity of this method is

quite high, (detection limit about 0.025 ppn ammonia), it is net an ideal one.

Ammonia also reacts with a derivative of pyrazolone called bispyrazolone

in the presence of an oxidising agent (e.g. chloramine T) at pH 6.0 to

produce rabazoic acid which can be extracted into trichloroethylene as the

(12)yellow undissociated acid The absorbance of the extract at 450 nm is

a linear function of ammonia concentration in the range 1 to 500 pg of

NH4-N per litre with a standard deviation of 0.9%. The difficulties of

this method are that reagents are not commercially available and extraction

is required. Also the conditions for the reaction are very critical and

it has side reaction to produce coloured compound (bispyrazolone blue) which

3

gives rise to error.

By using the. N-l -naph thylethy 1 enediamine and sulphardlamde reagents

to determine ammonia concentration? one has to oxidize ammonia into

nitrite by hypochlorite or by the enzyme nitrosomon a s e ur op ae a° Long time

is needed for the process. Also many substances vill give interferences,

such as: Amines, oxidising agents, reducing agents, complex forming agents,

precipitants, substances which disturb the acid-base balance, and coloured

substances etc.

o-(Benzenesulfonamido)-p-benzo quinone in benzene leacts with ammonia

in dioxane to produce a coloured compound with an absorption maximum at I48O

(14)nm . But the preparation of the reagent is tedious aid the sensitivity

is low. Also other primary, secondary, and tertiary amines can produce the

same colour and cause interference.

An onz,yraic assay of ammonia involves the reductive aminatron of 2-oxo—

glutaric acid (in the presence of NH, NADIi and glutamate dehydrogenase) to

glutamic acid and conversion of the NADH into NAD, followed by measurement

of the decrease in extinction at 30 nra; the ammonia content is determined

by reference to a calibration curve Since NADII is quite unstable and

expensive, the method is not generally used

Derivative spectroraetric method can be employed to determine ammonia

concentration by converting ammonium ion to ammonia; and determine the

vapor pressure by direct measurement of the absorbance of ammonia at 203'•

But it has the disadvantages that a closed system is required, and the

standard deviation is high.

Activation analysis has a very high sensitivity for nitrogen but the

4instruments are very expensive and it can only determine the total nitrogen

content. Some elements e.g. interfere with the analysis'.

The simplest method for detecting ammonia is the ion-selective electrode

(18)

method . Ammonia sensitive electrode is available commercially now (e.g.

Orion model 95-1Q). But the sensitivity is low and relative error is high

at low concentrations$ volatile amines also interfere and some ions (e.g.

Hg I) can damage the electrode.

The Berthelot reaction method (also named phenol-hypochlorite method.

or indoohenol blue method) seems to be the most successful procedure for the

determination of ammonia. It is simple9 reliable and seems to overcome

all the difficulties encountered by the Nes slier's method.

There are some difficulties in this method. The reagents used by

different workers were not the same and the employment of suitable catalyst

was not realized until recently. The experimental conditions were quite

(19)confusing. Aisc some inorganic and organic interferences had been reported'1

The purpose of this research is to make a detail study on the reaction in

order to suggest a successful procedure and propose a method to solve the

problem of interference.

5II.HISTORICAL REVIEW OF RERTHELOT REACTION METHOD

The reaction of ammonia with phenol and hypochlorite in an alkatns

mudium to from a blue coloure solution was first reported by Berthelot in

1859(20).The blue producat was believed to be closrly rolated to indophenol.

Tge reaction was rediscovered by Cotton(21)in 1874 and was applied by Thowas(22)

ti determine ammonia in biological fluids. Van Slyke and Hiller(23)used this

reaction to measure the ammonia content of blood by reacting the separated

ammonia with a 100 C solution of alkaline sodium phenatc and potassium hypo-

chlorite, they claimed that the lowest limit of detcction is 200 mg NH3-N per

litre(ppm).

The sensitivity of the Van Slyke procedure was increased three folde by

Russell(24),who eliminated the excess alkali nad added a small amount of

manganese salt cs a catalyst. In 1950.Kulenok(25)had recommended this

method for water analysis in preferense to Nessler's mothod. In 1953.

Frankenboug and others(26)aattempted to use this reaction to determine the

nitrogen content in tobacco, they found that the reaction was interfered by

the presence of nicotine. In the same year.Riley(19)studied the reaction

and reported that the blue colour would be more intense and Stable when the

temperature was decreased from 100 C to 70 C. the manganous concentration

was reduced to one third of that reportcd by Eussell, and the concentrations

of phenate and bypochlorite were doubled.

Crowther and Large in 1956(27)discovered acctone can act as a catalyst

for the Berthelot reaction. a significant increase in sensitiveity was observed,

and the reaction can be conducted at room temperature with maximum colour

development within 20 minutes.

6

The most widely used catalyst of the Berthelot reaction— sodium

nitroprusside (sodium ferrinitrosopentacyard.de) was first discovered by

(28)

Lubochinsky and Zalta in 19r but found to be greatly useful by Brown

and others in 19571 Mam30 found that sodium nitroprusside. was far-more

(31) (32)efficient than all other catalysts including manganous' argenuous 75

ferrousand(33) cuprous (32). It further increased the absorptivity of the

product compared vjith that produced by acetone catalyst.

In a detail study of the reaction using acetone and sodium nitroprusside

(34)as catalyst Harwood concluded that nitroprussj.de was the better catalyst

Since the colour produced by nitroprusside was more stable than thau by

acetone the time for colour development was shorter and the colour intensity

was less sensitive to fairly wide changes in hypochlorite concentrations.

which could have a marked effect when using acetone as the catalyst and

nitroprusside considerably increased the sensitivity as compared with acetone

It is of interest tc note that the use of oxidants other than sodium

hypochlorite has been reported Kalplin and Satsko reported the use of

(33)sodium hypo oromite and Bolleter et al- reported the use of saturated

chlorine water under slightly acidic conditions (in half saturated boric

acid and 08% phenol) In a qualitative study of indophenol formation

Soloway and Santoro had used a large number of oxidants including several

N-halocompounds persulfate and others Stegemann and Loescheke described

the use of chloramine-T (CE3C6H4-SO2NC1Na.3H2O) in a routine procedure

Although Weller' had reported the chloramine-T gives a much lower

absorptivity than sodium hypochlorite it has the advantage of simple

preparation in well-defined concentrations unlike the chlorine water or

hypochlorite and it is more stable than hypochlorite solution. But the most

7widely use one is still the sodium hypochlorite which can be purchased easily

and give higher sensitivity.

By now, the Berthelot reaction method has been applied in all fields

of chemistry, such as the analysis of ammonia in blood(38), fertilizer(39),

boiler feed-water(40), sea-water(41), polluted water(42), urine(43) and

polluted air(44). Also auto-analyzers based on this metod have been designed(34),

and applied to the routine analysis of ammonia.

8

III. MECHANISM OF BERTHELOT REACTION

1. Introduction

Berthelot in 1859 discovered that the reaction between ammonia

and phenol in the presence of hypochlorite formed a blue colour solution

but the nature of the reaction was not known at that time. Tarugi and

(45)Lanci in 1912 stated that an amine or imine in the presence of hypochlorite

re acted with phenol to give p ~b e na o qui no no xy ph e nyl i mine (i.e. indophenol)•

Although Lapin' had suggested a mecha i. vi- XiX for the reaction in 19lSky yet

(33)

it was not widely known Until 1961 Bolleter et al ov J proposed a

sequence of reaction steps of the Berthelot reaction and arose the interest

bo study the reaction. The steps he suggested were as following?

There are some evidences to approve the mechanism. Firstly it is

reasonable to believe that the first step is the reaction between ammonia

and hypochlorite to give chloramine% since this is a standard method for

(47)the preparation of chloramine

Secondly p-benzoquinonechloramine is believed to be an intermediate

of the reaction. It can couple easily with phenol to form the yellow

associated indophenol which gives intense blue colour in alkaline solution

9

Also p-aminophenol can be oxidized by hypochlorite to give the

(49)D-benzoQuinonechlorxmine

and it is well known that p-amnophenol an react with phenol to give

(48)indonhenol.

It further proves that p-bensoqudnonechlorimine is an intermediate of the

reaction

2.Catalyst effect

From above.? it is quite reasonable that the mechanism propsed by

Bolleter et ale is acceptable» But he did not use any catalyst at that

time. and the effect of catalyst on the reaction is quite interesting

Acetone increases the rate of formation and the amount of indophenol

Tetlow and Wilson proposed that its effect was bo increase the

concentration of one or more of the reactants in the rate determining

step by hinder or prevent one or more side reactions that would otherwise

decrease the concentration of the reactants Also the acetone catalyzes

the formation of monochlcroamine9 but the exact function of acetone is still

not known

Horn and Squire showed that sodium nitroprussids (sodium pentacyano™

nitrosyloferrate) reacted with sodium hydroxide to forra ferrous nitritopenta-

cyano complex in alkaline solution,

10

and said that this was the actual catalyst for the Bertheiot reaction

The catalytic effect of ferrous nitritopentacyano complex was also

confirmed in the present work. Sodium nitroprusside was added to solutions

without making alkali first. The rate of the reaction at the beginning was

very slow as shown in curve A of Figure 1. After the catalyst was alkalized

the reaction went rapidly If nitroprusside was alkalized first it would

change into nitritopentacyano complex the reaction would start immediately.

The curve for the rate of reaction was shown in curve B of Figure 1.

3. Mechanism

The mechanism of the reaction may be presented as followings

(1)

(2)

(3)

(4)

(5)

(6)

3.1 Step (3)

At step (3),NH2Cl can further react with OCl to form NHCl and NCl3(50)

NH3+ HOCI

NH2C1+ H0C1

NHCl2+ HOCI

NH2CI+ H2O

NHCL2+ H20

NGl3+ H20

11

Absorbance

0.6

0.4

0.2

0 5 1015

B

A

Time, min.

Fig. 1 : Gurves of the Berthelot reaction using nitroprusside and

pre-alkalized nitroprusside as catalyst.

(ammonia concentration = 0.5 ppm)

A: Rate of Berthelot reaction using nitroprusside (without

pre--alkali%ed) a,. catalyst, showing the slower ni tial rate.

B: Rate of Berthelot reaction using prealkalized nitroprusside

as catalyst.

12

these reactions are pH dependent• At pll h5 to 5»0 dicnlor amine was

the sole product Only mo no c Ho ramine (NHCl) was found at pH value above

85 neither dichloramine (M Gln) nor trichloramine (NCI,) were found

(52)under this condition

Prochaakovd' had shown that the mechanism of the reaction of ammonia

chlor amine T and bispyraaolohe to form, rub a so ic acid similar to Berthelot

(53)reaction was via the intermediate NKC10 and Newell suggested that the

actual molecule participated in Berthelot reaction might also be. dichloramine

The result in present work is contrast to his hypothesis• The Berthelot

reaction can be divided into two steps the first step is the reaction

between NH, and NaOOl or NHClp• In the second step the products of the

first step condense with phenol to form indophenol. By varying the pH of

the first step so as to produce only NHgCl or NIICl then let it to reset

with phenol at one final pH (pH 115)• From the amount of indophenol produced

the actual intermediate can be found• The pH effect in the first step on

absorbance is shown in Figure 2 It is found that when the pH is over 8%$

more indophenol is produced It can be concluded that the Berthelot reaction

is via the intermediate

(47)

NHgOl was prepared according to Inorganic Synthesis present

work. It can react with phenol solution to form indophenol without the

present of sodium nitroprusside. However the Berthelot reaction cannot

proceed at room temperature in the absence of a catalyst; It is quite

reasonable to believe that the function of sodium nitroprusside is to increase

the reaction rate of step (3).

13

Absorbance

1.2

0.8

0.4

4 6 8 1012

pH

Fig. 2 : Effect of pH value in the first step of Berthelot

reaction on the absorbance yield.

(ammonia concentration = l ppm)

14

Time, hours

Fig.3: Reaction rate of monochloramine arid phenol.

(About 0.002 M MIoCl and 1.7% phenol)

The reaction between mono chlo ram ne and phenol is pH dependents its

rates of reaction at different pH is shown in Figure 3

In the absence of a catalysts it requires very long time (over 5 hours)

for the completion of the reaction at pH value lower than 10. If the pH

value is higher than 12. the reaction can be completed in about 3 hours.

Compare to the rate of the Berthel'ot reaction (ammonia solution phenol+

hypochlorite) which is shwon in Fig.16 in Chapter VIj the reaction can

be completed in 10 minutes at pH over 12 it is much faster than that starts

from monochoramine. This may be due to the catalytic effect of sodium nl.tro-

prusside on the latter reaction step (i.e. step (b)).

15

3.2 Step(4)

Step (h) is not a single step in nature• There axe several possible

paths for this step

a pxAminopnenol as intermediate

It may pass through the intermediate p-aminophenol which is formed by

the reaction of monochioramine and phenol

then the p-aminophenol can react with hypochlorite to form p-benzoqinonechlorimine.?

which reacts quickly with another phenol molecule to form indophenol

b. p-Quinone as intermediate

p-Quinone may also be an intermediate of .bert helot reaction it can be

produced by the oxidation of phenol with sodium hypochlorite p-quinone reacts

with monochloramine to produce p-benzoquinonechlorimine and then couples with

phenol to form indophenol

16

c. Nitrene as intermediate

A nitrene intermediate is possibly involved in the reaction Transition

metal complexes can catalyse nitrene reactions by forming complexes with

(54)nitrenes and act as the intermediates in this kind ox reactions' So in

step {k)j nitroprusside may stablize a nitrene and make the reaction proceed

quia kly.

3.3 Step (5)

In step (5) the intermediate p-benz oquinone chlo rimine couples with phenol

and leads to the final product—- indophenol of the Berthlot reaction®

p-Benzoquinonechlorimine is a yellow crystal which can be obtained from the

reaction of p-aiuinophenol and sodium hypochlorite. This compound was

prepared in present work according to Beilsteins Handbuch dor Organ:,schen

(55)Chemie« The rate of its reaction with phenol by tracing the amount of

indophenol blue formed is shown in Figure ho This reaction is also dependent

on pH. At pH value over 10 the reaction proceeds very fast it can be

completed in 3 minutes or less® At pH 9«13 it requires more than 1 hour to

complete the reaction but it does not require sodium nitroprusside as

catalyst.

3.4 Step (6)

Step (6) is the ionization of indophenol in alkaline solution® Indophenol

is light yellow colour in acidic medium but shown intense blue colour in

acidic solution because under this condition it exists as ionized form and.

17

Absorbance

0.8

0.6

0.4

0.2

0 10 20 30

pH

10.96

10.23

9.13

Time, min.

Fig. 4 : Reaction rate of p-benzoquinonechloroimine and phenol at

different pH values.

(10 ppm p-benzoquinonechloroimine and 1.5% phenol)

has the resonance structures as following:

O N O O N O

They stabilize the molecule and change the absorption curve. The absorption

spectrum of indophenol blue is shown in Figure 5. Its absorption maximum

is at 634 nm.

18

Absorbance

0.6

0.4

0.2

400 500600

700

Wavelength, nm.

Fig. 5: Absorption spectra of indophenol solution.

(Absorption maximum at 634 nm)

(Ammonia concentration 0.5 ppm)

19

V. EXPERIMENTAL

1. Ness1 er's Method

1.1 Apparatus and instruments

A Brinkmann model 102 digital pH meter was used for pH measuremnts

absorbance measurements were done by means of a Hitachi 323 UV-YIS-NIR

recording spectrophotometer•

The block diagram of the spectrophotometer is shown in Figure 6 In

this spectrophotometer$ the wave-length can' be reset with accuracy better

than CU nm (in UV region). Tungsten lamp is used as the light source for

the visible range® The slit is automatically adjustable from 0.005 to 2 ram.

The accuracy of absorbance is better than 0.005 unit at 0J4 absorbance®

1.2 Reagents

a. Ammonia-free water

This reagent was orepared by redistilling distilled wafer. For every

litre of distilled water about 0.1 gram of potassium permanganate and 0.1 mi

of cone, sulfuric acid were added. Only the middle portion of the resistilled

water was collected. This redistilled water was used, to prepare solutions

in experiments.

b. Nessier reagent

(5)This reagent was prepared according to Yogel. About 35 g of potassium

iodide was dissolved in 100 ml of water and h% mercuric chloride solution

was added with stirring until a slight red precipitate appeared (about 325 ml

20

Slit Sector mirror

Light

sour.ce

Monochro-

mator

Sample

compartmentDetector

Iran smittance,

absorbance

conversion

circuit

Reference hold

circuit

Amplifier

Sample hold

circuit

Recorder

Electrical

connection

Optical path

Fig. 6 The block diagram of the Hitachi model 323 UV~VIS~HIR

recording spectrophotometer.

was required). Then a sodium hydroxide solution (120g25Qiril) was added and

made up to 1 litre with ammonia-fr ee water. A little more mercuric chloride

solution was added until there was a permanent turbidity- The mixture was

allowed to stand for one day and the sediment was removed by decantation.

The solution was kept in a dark-coloured bottle.

c. Standard, ammonium chloride solution

Trie 1000 ppm standard ammonia solution was prepared by dissolving 3cili g

of ammonium chloride (Merck GR) dried at 100 C in ammonia-free water and

21diluted to 1 litre. This stock solution was too concentrated lor most purposes.

The 10 ppm standard ammonia solution was made by diluting 10 ml of this stock

solution to 1 litre with ammonia-free water. All other concentrations of

ammonia solutions were prepared by diluting the stock solution in proper ratio.

1.3 Procedure

To each 50 ml of sample solution. 1 ml Nessler reagent was added; thoroughly

mixed and allowed to stand for 20 minutes then the absorbance was measured

at klO nia« For ammonia concentration in the range 0.1 to 1 ppm; longer time

(30 minutes) was needed for standing and the absorbance was measured using a

4 cm cell. For ammonia concentration in the range 10 to 20 ppm; the abscrbancs

was measured at wavelength 525 nm using a 1 cm cell.

2.- Direct Berthelot Reaction Method

2.1 Reagents

a. Acetone (Merck; OR),

be Ethanolic phenol solution (30%)

All phenol used in present work was prepared by vacuum distillation of

phenol crystals (Merck; GH) to ensure purity. The 30% phenol solution in

ethanol was prepared by adding 30 g distilled phenol to absolute ethanol and

diluted to 100 ml.

c. Sodium nitroprusside solution (025$)

0.25 g sodium nitroprusside (Merck GR) was dissolved in ammonia-free

water and diluted to 100 ml.

22

d. Sodium hypochlorite solution (BDH LR)

The hypochlorite content was found to be 6.02% by iodometric titration.

e. Solution A

It consists of 1.1 sodium hypochlorite and bio sodium hydroxide.- Proper

amounts of sodium hypochlorite, and sodium hydroxide were dissolved in ammonia-

free water and brought to 100 ml to make up this reagent.

f. Solution B

It consists of 37h% phenol and 0088$ nitroprusside. It was prepared

by mixing proper amounts of sodium nitroprusside solution and ethanolic phenol

solution in a volumetric flask and diluted to the mark with absolute ethanol.

All above solutions should be kept in amber bottles and placed in a

refrigerator but allowed to reach room temperature before using.

g Sodium hydroxide solution (5 N)

100 g sodium hydroxide- (Merck GR) vras dissolved in ammonia-free water

and made up to 250 ml.

h. Sodium citrate solution (10%)

Proper amount of tri~sodium citrate 5% hydrate (Merck GR) was dissolved

in ammonia-free water to make up this solution.

i. Buffer

Buffers ot different pH were prepared according to Handbook of Chemistry

and Physics. For pH 11.30 buffer it was prepared by mixing 50 ml 0.05 N

23disodium hydrogen phosphate dihyrate solution (Perking1s- Reagent AR) with 150 ml

0.1 N sodium hydroxyl solution.

j. Buffer plus completing agent solution

10% sodium citrate solution and pH 11.30 buffer were prepared in the

same bottle.

2.2 Procedures

a. Optimum conditions for Berthelot reaction:

To find the optimum concentrations of each reagent participated in

Berthelot reaction the concentrations of ail reagents were kept constant ex-

cept varying the concentration of the reagent being studied. By measuring the

absorbance of the final product the optimum concentration for that reagent.

was determined from the maximum absorbance.

For example to find the most suitable concentration for phenol 0.8 ml

0.25% sodium nitroprusside solution O.m ml 6.02$ sodium hypochlorite solution

and 2 ml 5 N sodium hydroxide solution were added, successively tc every 5'0 ml

0.5 ppm ammonia solutions. Then various amounts of 30% ethanoiic phenol

solution was added to make the final phenol content from 0% to 2% The mixtures

were shaken and allowed, to stand for 20 minutes before measurement at 63h nm

for indophenol. The optimum conditions of other reagents were determined

similarly.

b. Determination of sample solution

To 50 ml ammonia sample solution 5 ml buffer plus completing agent

solution 2o ml solution A and 2.$ ml solution B were successively added

24mixing thoroughly after each addition. The colour was allowed to develop at

room temperature (20- 25°C) for 20 minutes and the absorbanee was measured

at 63 k nm. For 0.1 to 1 ppm ammonia concentration.- 1 cm cell was used and

for 0.01 to 0.1 ppm ammonia concentration h cm cell was used.

All glasswares used were cleaned by washing initially with dilute

hydrochloric acid and rinsed thoroughly with ammonia-free water.

c. Interferences study

A series of solutions each containing 0.5 ppm ammonia and 100 ppm different

foreign ions were prepared. Then 5 ml buffer plus complexing agent solution

25 ml solution A and 2.5 ml solution B were successively added and the mixtures

were shaken to make the reagents mix thoroughly. The colour was allowed to

develop for 20 minutes and the absorbances were recorded at 63 k nm. The

absorbances were compared witn that of 0.5 ppm pure ammonia solution the

effects of the foreign ions were found.

3. Simple Diffusion Method

3.1 Reagents

a Sodium bisulfate solution (0.01 M)

13.808 g sodium bisulf ate (Mallinckrodt AR) was dissolved in ammonia-

free water and made up to 1 litre the solution would be 0.1 M in sodium

bisulfate. 0.01 M sodium bisulfate solution could be made from diluting this

solution 10 times.

b. Hydrochloric acid (0.01 M)

Cone. hydrochloric acid (Merck. GR) was diluted to 0.01 N by ammonia-

free water.

25

3.2 Apparatus

The simple diffusion flask used in the simple diffusion method for

determining ammonia concentration in solution was modified from an 250 ml

Erlenmeyer flask. The set up of the flask is shown in Figure 7 It consisted

of a flask and a collecting tube. A side arm (a) was used for the injection

of alkali. A needle puncture stopper (B) was used to cover the side arm the

diaphragm of the stopper can be punctured with a syringe needle and it will

sealed automatically after the needle is withdrawn. Three indentations(D)

were made at the neck of the flask for holding collection tubes The collection

tube (0) was made from a test tube or glass tube which could fit into the

neck cf the Erlenmeyer' flask. A rubber stopper. (E) was used to seal the mouth

Of the flask.

3.3 Procedures

a. Operation of the simple diffusion flask

After sample solution was- pitpetted into the flask the collection tube

was placed at its position with the absorbent and the flask was closed firmly

with the stoppers. Then it was put into a water-bath at constant temperature;

and alkali was injected by a needle syringe. The evolved ammonia was absorbed

by the absorbent. After a fixed time the collection tube was taken out and

the solution in it was tested for ammonia by direct Berthelot reaction method.

b. Optimum conditions for simple diffusion method

To determine the optimum conditions of the simple diffusion method all

factors were kept at constant values except the factor being studied optimum

condition of that factor could be determined from the occurrence of maximum

26

E

D

C

B

F

G

A

Fig. 7 : Apparatus of simple distillation method

for ammonia determination.

A: Side arm for injection of alkali.

B: Needle puncture stopper.

C: Collection tube.

D: 3 indentations for holding collection tube.

E: Rubber stopper.

F: Absorbent.

G: Sample solution.

27

abs or banco of the coloured product solution..

For example to choose the most suitable absorbent 5 ml 05 ppm ammonia

solution was put at the bottom of the simple diffusion flask Then 5 ml of

ammonia-free water wan pipetted into the collection tube the 11 ask was sealed

with the stoppers• It was put into water-bath at 30°C then 5 ml 5 N NaOH

was injected with a syringe. After a certain time the collection tube was

taken out. 0.25 ml Solution A 025 ml solution B and Co ml buffer solution

were added successively thoroughly mixed and allowed to stand for 20 minutes.

Absorbance at 634 nm was measured.

The concentration of the most efficient absorbent could be found by the

shortest time required to give quantitative absorption of the evolved

Similar procedures were employed to determine the optimum conditions of other

factors.

c. Determination of ammonium sample solution

The conditions of diffusion for£ .il sample solution were using 5 mil 0.01M-

sodium bisulfate solution as absorbent injection of 5 ml$ N NaOH and standing

in 50°C water bath for k hours. Then to the solution in the collection tube

which had. absorbed the evolved ammonia 0.25 ml solution A 0.25 ml solution B

and 05 ml buffer solution were added. The solution was mixed thoroughly and

the absorbance at 63k nm was measured after 20 minutes standing

U Application- Determination of ammonia content in the waste water of

The Chinese University of Hong Kong.

28

4.1 Reagents

All reagents used in Nessler's method, direct Berthelot reaction method

and simple diffusion method were employed.

4.2 Procedure

The waste water of CUHK before sewage treatment was diluted 100 times

before using the above methods, but the water after sewage treatment was

examied directly for ammonia concentration. The procedure were just the same

as those stated in the procedures for sample determination part of the Nessler's

method, direct Berthelot reaction method and simple diffusion method.

29

V. RESULTS AND DISCUSSION

1. Ness1er's Method

Hie calibration curves for ammonia determination using Nessler's method

were shown in Figures 8,9, and. 10. The concentration of unknown ammonia

solutions can be found by comparing the absorbance of the Nesslerized products

in the samples with the calibration curves. Hie calibration curves should be

checked frequently to assure the reliability of the measurements.

From Figure 8, it can be seen that in the absence of interference,

ammonia concentrations in the range of 1 to 10 ppm can be determined satis-

factorily by the Nessler•s method, with the procedure stated in experimental

part. Hie calibration curves were treated by least squares fitting, and the

correlation coefficient of this calibration curve is 0.988c The average root

mean square error is+ 02 ppm for 1 to 10 ppm ammonia concentrations» Hie

deviation may be due to the onset of turbidity at high concentration

af ter 20 minutes standing For more concentrated ammonia solutions direct

dilution of the Nesslerized product cannot be used, because it is colloidal

in nature, and the solution is inhomogeneous. However, ammonia concentrations

in the range of 10 to 20 ppm can be determined at the wavelength of 523 nm,

(Figure 9) There will be some sacrifice iu accuracy because of the lower sen¬

sitivity and the easy formation of turbid material, the correlation coeffi¬

cient of the calibration curve is only 0727$ that means the uncertainty may

be very high for the unkown ammonia concentration. Under the same experimental

condition, ammonia concentration in the range of 0.1 to 1 ppm can be measured

using a k cm. cell, a calibration curve is shown in Figure 'SO, the correlation

coefficient of this curve is 0.991$ and the average root mean square error is

30

Absorbance

1.2

1.0

0.8

0.6

0.4

0.2

0 2 46

810

NH3 conc., ppm

Fig. 8 : Calibration curve for the Nessler's method.

(wavelength at 410 nm, using 1 cm cell)

31

Absorbance

0.6

0.5

0.4

0.3

0.2

0.1

0 12 14 16 18 20

NH3 conc., ppm

Fig. 9 : Calibration curve for the Nessler's method.

(wavelength at 525 nm, using 1 cm cell)

32

Absorbance

1.0

0.8

0.6

0.4

0.2

0 0.2 0.4 0.6 0.8 1.0

NH3 conc., ppm

Fig. 10 : Calibration curve for the Nessler's method.

(wavelength at 410 nm, using 4 cm cell)

33

+ 0.02 ppm in this concentration range.

The spectra of the Nessler's reaction product is shown in Figure 11.

Since it is colloidal in nature, and it does not have an sbsorption maximum

in the visible ranga, this limits the accuracy of Nessler's method. It's

stability is low, turbidity occurs after half an hour standing. Numerous

factors cause erratic results with the method and numerous compounds interfere

with the reaction as stated in Chapter I. All these are the shortcomings of

Nessler's method for the determination of ammonia.

Absorbance

2

1

340 400 500 600 700

A B

Wavelength, nm.

Fig. 11 : Absorption spectra of Nessler's reaction product.

(A: 2 ppm ammonia soln.; B: 20 ppm ammonia soln.)

34

2e Direct Berthclot Reaction Method

Owing to the contradiction of the conditions for the Berthelot reaction

(34)method proposed by other workers, all factors vhich might affect the ab-

sorbance of the coloured product were re-examined. These included preparation

of reagents, concentration of reagents, temperature, alkalinity, and. time

factor.

The reagents for Berthelot reaction basically consist of phenol and an

oxidant, also a catalyst is needed for the reaction to proceed at room tempe¬

rature. In a basic medium of high alkalinity, the reacting species, actually

is phenate ion, but due to the instability of the sodium phenate solution,

vhich has to be renewed for every two days because of the darkening of the

(27)reagent, a phenol solution is prefer-ecT

For the oxidant, reagent's other than sodium hypochlorite had been reported

as mentioned earlier in Chapter II, but the results are not as good as sodium

hypochlorite.Therefore sodium hypochlorite was used throughout the expert-

raents

A number of compounds have been tried as catalyst in present work. The

maximum absorbance and time required are listed, in Table 1. Results show

that nitroprusside has the highest sensitivity, and the time required for

full colour development is also the shortest among al3. compounds» So it was used

as the single catalyst.

Since the solution is very alkali, coraplexing agents had been used to

(57)prevent the formation of metal hydroxy1 precipitates« CDTA (cyclohexyl-

(58)

trans.-1,2-diaminetetraacetic acid) had been proposed by Roskam, but it

is not widely used. Tine more common chelating agent—KDTA (ethylene-

(59,30)diaminetetraacetic acid)—had been used by some workers, but it was found

35

Table 1

The catalytic effect of some compounds on Berthelot reaction.

Substance Molecular

formular

Absorbance Time re-

quired

(min.)

sodium nitro-

prusside

NaFe(CN) NO5 0.520 20

acetone CH. COCH3 3

0.208 30

patassium ferro-

cyanideK Fe(CN)

4 60.403 10 hours

potassium ferri-

cyanide'K3Fe(CN)6

potassium cobalti-

cyanide

K Co (CN)3 6

potassium cobalti-

nitriteK Co(N0 )

3 2 bproduce no

blue colour

potassium chloride

barium chloride

disodium tetra-

borate

potassium sodium

(+)-tartrate

KCI

BeCI

Na B O .1OH O2 4 7 2

KNaC H 0 .H O4 4 6 2

Absorbances were obtained from 0„5 PPm NH and 001%' substances solutions.

36

in present work that no blue colour- was formed Citrate ion also forme stable

#

complexes with calcium and magnesium ions , and it does not lower the colour

(61)intensity of the reaction product. So it was chosen as the completing

agent in present experiments,. The concentration of the complexing agent was

chosen at VL The effect of the complexing agents on the colour intensity

is listed in Table 2

Table 2

The effect of complexing agent on the colouration of

of Berthelct reaction

complexing agent absorbance

blank

1% sodium citrate

2% sodium citrate

1% jUDTA

2% EDTA

0.514

0.520

0.512

0.012

0.000

Stability cons'cants on the metal citrates (60).

log BMHLlog BML

Calcium

Magnesiun

8.02+ 0.15

7.66+ 0.19

3.50+0.06

3.38+ 0.07

B is the stability constant of MHL complex andMHL

B is the stability constant of ML complex.MHL

(N= metal ions; L= citrate ligand)

37

(62)For the combination of reagents 9 Weather burn had shown that phenol

plus nitroprusside, alkali plus hypochlorite would give results with greatest

precision.Since only two solutions are required in that combination, and

both of them are reasonably stable, it simplifies the analysis procedure®

For the sequence of adding the solutions, some preferred to add phenol

(19.63,64)before hypochlorite» But in present.work, results showed that hypo-

chlorite-phenol sequence can give higher' absorptivity (see Table 3).

Table 3

Effect of the sequence of adding phenol and hypochlorite

in.Berthelot reaction

Absorbance

Ammonia

c once ntration

Hvpochlorite, then

phenol

Phenol, then

hypochlorite

05 PP

1.0 ppm

0.513

1.024

0.38c

0.768

Die reason may be that if phenol is added first, the later added hypo-

chlorite may react with phenol. Thus reduces the amount of chloramine which

(33). will react with phenol to produce indophenol» Also j.f the interval between

adding the two reagents is too long, decrease in absorbance will be observed,

as shown in Table k, so it is better to make the interval as short as possible.

2.1 Determination- of optimum conditions

A series of experiments were undertaken to obtain the optimum conditions

38Table 4

Effect of the time interval between hypochlorite

and phenol in the absorbance.

Time interval

(min.)

Absorbance

0 0.520

2 0.500

5 0.470

7 0.470

100.435

15 0.412

of the Berthelot reaction method.

a. Effect of phenol concentration

By keeping the concentrations of all other reagents at constant valures,

only the phenol concentration varied, the effect of phenol concentration

on the Berthelot reaction would be observed. By measuring the maximum absor-

bance of the colour, it was found that the absorbance rised quickly when

phenol concentration exceeded 0.6% as shown in Figure 12. Hence the 1.8% of

phenol concentration was chosen as the optimum value for safe reason, and

was used throughout the experiments.

b. Effect of sodium hypochlorite concentration

A peak occurred at the hypochlorite concentration of 0.042%, and the

39

Absorbance

0.8

0.6

0.4

0.2

0 1 2

Phenol cone., %

Fig. 12 : Effect of phenol concentration on Berthelot reaction.

(ammonia concentration = 0.5 ppm)

Absorbance

0.8

0.6

0.4

0.2

0 0.02 0.04 0.06 0.080.1

Sodium hypochlorite, %

Fig. 13 : Effect of the amount of sodium hypochlorite on Berthelot

reaction.

(ammonia concentration = 0.5 ppm)

40absorbance decreased when hypochlorite concentration was- further increased

(Figure 13) It does not agree with the results of some other workers'

They1 said that a plateau was observed for a vide range of hypochlorite con¬

centrations, and fading occurred only at very high hypochlorite concentrations

It may be that the reverse sequence of addition of phenol and hypochlorite is

the reason for the difference So that the concentration of hypochlorite

sVirYnl d Hp wyofnT 1 v ArmhrvVI 1 Ad.

c. Effect of Sodium Nitroprusside

Hie colour intensity of nitroprusside is quite high when it is changed

(38)to ferrous nitritopentacyano complex in alkali solution-J° 5 it will mask the

indophenol blue colour when high concentration of catalyst is used® So mini¬

mum amount of nitroprusside is prefered for maximum sensitivity of the

Ber-theiot reaction method From Figure 1 k, the absorbance of the product

reaches maximum value in the range from 0.005% to 002%.The 000-}% vas choser

as the recommended value} of which the concentration is quite low and the

colour interferenceis negligible.

d, Effect of pH

pH has a great effect on the intensity of the indophenol blue colour

(34)It also affects the rate of colour devlopment and colour stability

Since the'colour product is reddish brown in molecular form, and only ionizes

in alkaline solution to give intense blue colour, so the final solution

must be highly alkali

The effect of pH value on the colour- intensity shows that maximum coloui

can be obtained between the pH values of 11.0 and 120 (see Figure 15).

(30,34)These data are in agreement with other-workers

41

Absorbance

0.6

0.4

0.2

0 0.005 0.01

Sodium nitroprusside conc., %

Fig. 14 : Effect of the amount of sodium nitroprusside on Berthelot

reaction.

(ammonia concentration = 0.5 ppm)

Absorbance

0.6

0.4

0.2

89

10 11 12 13

pH

Fig. 15 : Effect of pH on Berthelot reaction.

(ammonia concentration = 0.5 ppm)

42

e. Tims for Colour Development

There is a large difference in opinions concerning the time required

for maximum colour development Typical time ranges from a few minutes to

(34)more than an hour. Experimental results show that the time is pH depen-

dent, the higher the pH value, the more rapid the rate of colour development«

At pH 10.30, maximum colour cannot be obtained in more than 1 hour standing

after the reactants have been mixed For pH above 1130? only 20 minutes is

required to obtain the maximum colour. And'even shorter time (about 10

minutes) is required for pH higher than 12.00, but the colour tends to fade

away after reaching the maximum value (Figure 16).

Time min

Fig. 16: Rate of colour development at different pH values

of Berthelot reaction.

(ammonia concentration= 0.5 ppm)

43Therefore the pK value of 1150 is recommended and one can measure the

absorbance of the sanrole 20 minutes after the reaction, and results shew that

the colour is stable for another 25 hours.

f. Effect of Temperature

All above experiments are conducted at room temperature (about 20 C)0

Since the time required for maximum colour development is reasonable, no

detail study of temperature effect on colour development has been made.

However, on rough observations, no marked effect was noted on the reaction

at 50 C temperature in a water-bath.

g. Conclusion

For direct Berthslot reaction method, the concentrations of reagents for

obtaining the optimum sensitivity are as follow:

Phenol

Sodium nitroprusside

Sodium hypochlorite

pH

: 1.8

: 0.005%

: 0.0042%

: 11.50

2.2 Calibration curves

In the absence of interference, ammonia concentration in the range from

0.1 to 1 pprn can be detected sucessfully with a 1 cm cell by the direct

Berthelot reaction method using the optimum conditions described above.

The calibration curve is shown in Figure 17? the correlation coefficient

of this curve is 0.999. and the root mean square error is 0.005ppm so it

follows the Beer's law quite closely. The concentrations of unknown

44

1.0

0.8

0.6

0.4

0.2

0 0.2 0.4 0.6 0.8 1.0

NH3 conc., ppm

Fig. 17 : Calibration curve for direct Berthelot reaction method.

(wavelength at 634 nm, using 1 cm cell)

45

Absorbance

0.8

0.6

0.4

0.2

00.02 0.04 0.06 0.08

0.1

NH3conc., ppm

Fig. 18 : Calibration curve for direct Berthelot reaction method.

(wavelength at 634 nm, using 4 cm cell)

45

ammonia solut ion; can be dctermiir d by direct: rending from th s colli ration

curve.

In more dilute solutions, longer cells should be used, th- calibration

curve of using cm ccO.l for the ammonia concentrations fron 0.01 to O.'l upr.i

is shown in Figure 18, The correlation coefficient of this curve is

and its root mean square error is 0.000'f ppm for this concentration range.

So it is quite confident to use the Ferthclot reaction method for the deter¬

mination of 0.01 to 0.1 ppm ammonia.

2.3 Intorference of the Boy'helot re-cticn

a. Inorganic .interference

(10)In 1933, Riley' reported that magnesium ion interferes with the

Berthelot reaction by forming hydroxide precipitate, and made the method not

usable for direct determination of ammonia in sea water. This problem was

(61)solved by adding complexing agents (eng. soduim citrate) to the solution''•

But some other inorganic ions still interfere with the reaction.

In present work, the interference effects of some common inorganic ions

that may exist in water were studied. The results are lifted in Table 3.

The concentrations of the foreign ions in the present work was at 100 ppm

level, in a solution containing 0.3 ppm ammonia. By comparing the absorbcaice

of the solution after addition of the reagents to en interference free ammonia

solution, the deviation may be considered to be caused by that foreign ion.

For magnesium ar.d calcium ions, since they can form stable complexes

vith sodium citrate, they will not interfere with the reaction. Aluminium

and zinc ions form precipitates in the alkali solution if sodium citrate is

absent. Results show that their interferences are quite small in present

47

Table 5

Interference study of inorganic ions on

Direct Berthelot Reaction Method

(100 ppm in 0.5 ppni ammonia solution)

Substance Added as Concentration Absorbance V id, J- O iT

(ppm)

aluminum

calcium

cobalt

copper(II)

ironCII)

iron(III)

magnesium

manganese

mercury

potassium

sodium

sulfite ioi

zinc

A1C13

CaCl0

CoC1A

CuSO. 5H O

FeCj 4H2O

FeCl .OH O3 2

NgC12

MnCl2 4H2O

HgCl.2

KC1

HaCl

NaSO2 3

ZnCl

100

100

100

100

100100

100

100

100

100

100

100

100

100

0.500

0.504

0.281

0.350

0.306a

O.399a

0.515

0.326b

0.368a

0.506

0.502

0.337

O.499

-C.010

-0.006

-O.200

-0.160

-0.204

-0.111

0.005

-0.184

-0.142

-0.004

-0.003

-0.173

-0.011

deviation= absorbance of the solution v:ith foreign compounds

- 0.510 (absorbance of the solution with 0.5 ppm ammonia only),

athe solutions were green in colour.b

brown precipitate appeared., and solution was groan in colour.

48

method and small amounts of thorn in cample solutions are tolerable, bodium

and potassium ions are also among those ions that do not interfere rich the

reaction.

Copper ion is blue colour in nature and should increase the absorption

value at 63 nm, but results show that absorption value is actually decreased,

Ti-at means the interfering effect of copper ion is high. Cobalt ion also

lowers the absorbane of the solution, and its effect is larger than copper

Manganese ion forms manganese (IV) oxide precipitate when sodium hypo-

chlorite is added to the solution. It also makes the final solution green

in colour, co the presence of a small amount of the manganese ion will cause

a great error.

Iron(ll), iron(lli) and mercury ions also produce green coloured products.

with the reactants. The nature of these products are unknown, they may be

the derivatives of indophenol.

Ordinary anions seem to cause no interference with this reaction. Gnly

(58)sulfite ion has interference, and the result is consistent with other worker

b. Organic interference

For organic compounds, Vearne found that some nitrogenous compound

also gave blue colouration with the hypochlorite-phenol reaction. Several

nitrogen containing organic compounds and hydroquinone have been studied foi

their interfering effect on Berthelot reaction in the present work, and the

results are listed in Table 6.

Urea interferes vith the reaction seriously as shown in the table.

(65)Wearne stated that it can yield precisely the same magnitude in absorbanco

of the corresponding ammonia solution. This means that, it can replace the

49

Table 6

Interference study of organic compounds on

Direct Berthelot Reaction Method..

(100 ppm in 0.5 ppm ammonia solution)

Substance Concontration

(ppn)

Abcorbance deviation

alanine

p-aminophenol

aniline

aspartic acid

n-butyl amine

sec-butyl amine

tero-bucyl amine

glutamine

glycine

hydrazine

hydroqoinone

hydroxylamine

sulphanila mide

urea

unic acid

100

100

100

100

100

100

100

100

100

100

100

100

100

100

100

0.510

2.040

3

0.530a

1.400

0.105b

0.165b

0.505

0.940

0.806

0.220c

O.825a

1.220

1.880

0.778

-0.005

1.455

2.4

0.015

0.885

-0.410

-0.350

-O.010

0.425

0.345

-0.295

-0.230

0.705

1.365

0.265

deviation= absorbance of the solution with foreign compound

- 0.515 (absorbance of the solution with 0.5 ppm ammonia on1a

colourless crystals precipitate ou

b light green in colour.

c deep brown in colour.

d green in colour.

50position of ammonia to produce indophenol and c use error.

Uric acid, sulphanilamidc and hydrazine also cause positive deviation

to the reaction, they may produce indophenol or related compounds with tlie

reactants, but their interfering poser is lower than that of urea.

The interfering effects are quite different among amino acids. For

the several amino acids studied, glycine increased the absorption value

largely, but a la mine, a spar tic acid and glutan.ine had little- effect on the

absorbance.

Aminos also interfere with the reaction.. n-Butvl amine, one of the

primary amines raises the absorbance to a very high value, but secondary

and tertiary amines, e.g. sc-o-butyl amine and tort-butyl amine, lever the

absorption value. For aromatic amines, aniline gives an extraordinary high

absorbance since it is easily oxidized by hypochlorite to produce intense

coloured procucts.

Ihe colour of the product, with the,presence of hydroxy famine is green,

(33)it is consistent with the work of Bolleter, he found that it made no

blue colour when hydroxylamine was present in the solution, jo-Amincphenol

can react with phenol in alkaline medium to give indophenol, so it interferes

seriously with the Eerthelot reaction method. Hydroquincne produces only

deep brown coloured solution if it is present in the sample solution, it

may be the oxidation product of hydroauinone with hypochlorite.

Therefore, the interference of organic compounds is much larger than

that of inorganic compounds. Fortunately, these nitrogenous compounds are

seldomly presented in natural water, the direct Berthelot reaction method

can still be applied to certain sample solutions.

51

2.4 Conclusion

Comparing the results of Nessler's method to Bertholot reaction method-,

it is found that the latter one is superior in many ways to the farmer o..„

Firstly, the product of Nessler's reaction is a colloidal compound nh:.ad-

gives no absorption maximum in the visible region. The coloured complex of

Bert-helot reaction is a truely soluble compound, and it has an absorption

maximum at 634 nm.

Secondly, the experimental conditions of Bertheiot reaction are less

critical than that of Nesslerization.

Thirdly, the sensitivity of Bertheiot reaction is higher than that cf

Nessler's method, the detection limit is at least 10 times .1 o:er.

Fourthly, there are less interferences from foreign compounds to the

Bertheiot reaction method.

Fifthly, the product of Bertheiot reaction is far more stable than that

of Nessler's method-, it remains stable for more than 21 hours with no change

in absorbance, but that of Nessler's method is only stable for about half an

hour

In general, to determine ammonia by the Nessler's method can be replaced

by Bertheiot reaction method. Direct ammonia detection can be done for sample-

solutions in many cases where the Nessler's method is not applicable.

52

3. SinnfLe Diffu xion Mot;hcd

Since the direct Berthelot reaction method is interfered by a number

of inorganic and organic compounds even in the presence of complexing agents,

method must be sought to solve this problem, The method should be simple,

accurateand give lew detection limit. It is best to be an in situ detection,

because ammonia in sample solution easily loses to atmosphere and large

error will occur if the sample has been stored for a long time.

A simple method for the extraction of ammonia may be done by diffusion.

Ammonia Ccn easily evolve from solutions by the addition of a strong alkali.

It is highly soluble in vat or. especially in acid solutions. So after

absorbing the ammonia in an absorbent, its concentration may be determined

by direct Berthelo': reaction method. In this way, interfering compounds

in the original sample will be largely eliminated.

An apparatus based on diffusion principle was designed, its shape and

function ver discussed in Chapter IV.

3.1 Determination of optiman conditions

A series of experiments vere undertaken to determine the optimum

conditions using this apparatus for ammonia determination.

a. Mature of the absorbents

Ammonia is basic in nature, an acidic solution will absorb larger

amount of ammonia than a neutral one« Water, hydrochloric acid and sodium

bisulfate have been tested as the absorbents. Kesults show that sodium

bisulfate is the best absorbent. It requires a shorter time for absorption.

Also a dilute absorbent is prefered, otherwise large amount of alkali will

53

be needed to neutralized it. Feci: Table, it can be ft-on tb .t the be, I

concentration of acidic absorbent 0.01 M Naif.:ojt and it wax; used an the

absorbent throuwhout the e:reri meni;sr

Table 7

Effect of the concentration of absorbent (NaII30)„

Oimmonia concentration= 0.5 pen)

HaHSO4 Cono.4

(H)

1 0.5 0.1 0.01 0.005

Absonbance 0.505 0.486

No blue colour formed due to tee acilic condition for normal

procedurer

b. Volume of the absorbent

The total amount of ammonia evolved vcniaun the same for a. fixed

amount cf a given sample solution, the smaller the volume of the absorbents,

the higher the ammonia concentration '.ill be in the final solution. A

larger amount of absorbent would not increase the rate of absorption

because the surface area cf the solution Inside the collecting tube remains

the same. A smaller absorbent volume can lower the detection limit of

ammonia, but if it is too small, it will be more difficult to handle. The

smallest volume of solution required for absorban.ee measurement using 1 cm

call to be 5 ml, and was uoud throughout the expertinerts.

c. Volume of sample solution.

Experimental results show that the rate of absorption of ammonia is

54

very much dependent on the sample volume. IT the volume of so mole .solution';

is too large, intolerable time will be required to complete the absorption

process, and if the volume of sample solution is small, shorter time- will

be required. For b nil of sample solution in which b ml of 5 N haOH was

added the diffusion process would be completed in hours at 10 C. The

absorption process of different amounts of sample solution is shown in

Figure 19.

d. Amount of sodium hydroxide

About half of the ammonium ions change into ammonia at pH nd

995 t pH 11.6v;, so that at higher pii values, ammonia will be evolved

much easier. But pH effect will be small if pH of the solution is over 11.6.

In fact, there is still another effect on the rate of ammonia evolution,

The evolution tension of ammonia in the solution can be increased by the

(67).presence of large amount of foreign iors. In present work excess emeu; f

of sodium hydroxide was added to the sample solution to speed up the rate of

absorption process. The effect of the amount of sodium hydroxide on the

absorbance is shown in Figure 20. For 5 ml of sample solution, b ml of b 1T

sodium hydroxide solution seems to be adequate, and it is used throughout

the experiments.

(19)The alkali employed by Riley' in his distillation method was pH 9-lb

buffer solution, but is was found to be not so effective as the b hT sodium

hydroxide in present work. Potassium carbonate had been found to have the

most powerful foreign ions effect on ammonia evolution. However, it has

been tested that under the conditions employed in present work, sodium

hydroxide was as effective as potassium carbonate,, Therefore, the most common

Alkali—sodium hydroxidc—was comployed.

55

Time, hours

Fig. 19: Effect of sample volume on ammonia absorption rate

(5 ml 05 ppn NIL solrw: 5 ml 0..01 M NallSO, as absorbent;

5 ml 5 N NaOII as alkali; a'.50°C.

(10ml 05 ppm NH soln.:3

10 ml 0.01 N NaHSO, as absorbenr;4

10 ml 5 N Ha OH as alkali; at 50°C.

(50ml 0.5 ppm NH3 soln.: 10 ml C.01 M NaHSO. as absorbent;4

50 ml 5 N Ha Oil as alkali; at 50°C.)

56

5 N NaOil, mj

Fig. 20: Effect of amount of 5 EaCh on absorbance,,

(5 ml 0,5 pom ammonia so In.; 5 nil 0,01 M MallOC as absorbent;

at 50°C foi A hours.)

e„ Temperature dependence, of the diffusion process

The higher the temperature, th° quicker will the diffusion process be

completed. When the temperature was very high (i.e. over 60 C), the

pressure of gases inside the flask vould be too high and the stoppers

might be blown off, hence to increase the chance of leaking. Also at such

high temperatures, it might hydrolyze the nitrogenous compounds into ammonia

and cause errors. The absorption vould be very slow if the temperature was

too low. These tvo factors should be considered carefully.

The absorption process for 5 nil sample solution and 5 nil 5 N sodium

hydroxide can be completed in A hours, without the leakage of ammonia gas.

57

Only about So j of ammonia was absorbed at 'fO C after h hoiu, and sti]1 Van

at room temperature. So the absorption process must be carried out in

water-bath to ensure it can be completed in a reasonable time, and['') C is

the recommended temperature.

f. Effect of stirring

Stirring can increase the rate of ammonia, evolution, but the effect

was found to be small in present work. The timt required lor the c..' die Lion

of the absorption process at room temperature with stirring was still vary

long. And it had little effect when the temperature was SO C or above.

Also it was very tedious in carrying out thi s procedure even using a

magnetic stirrer, and made this simple method complicated. So stirring vas

not used in the present work.

,g. Time requirement for the diffusion method

All the above factors hav some effects cn the rate of diffusion pro¬

cess,. Absorption process can be achieved in shortest time by using smallest

volume of sample solution, excess amount of sodium hydroxide and highest

temperature. From Figure 19 and 20, it can be found that with the

present apparatus, using p ml sample solution, ml 5 bT sodium hydroxide and

90°C temperature, the time required for completing the absorption process is

h hours.

h. Conclusior

The optimum conditions of the simple diffusion method using the piesent

apparatus are listed below:

58

Sample volume: 5 ml

Absorbent

Alkali

Temperature

Time required

5 ml 0.01 M NallSO.4

5 ml 5 N NaOH

50°C

h hours

It should be noted that the optimum conditions will be changed it' a

differently designed diffusion apparatus is used. And if there is another

method which can determine trace amount of ammonia in very small volume,

e.g. 1 ml of less, time requirement can be shortened by reducing the volume

of sample solution without the dilution effect.

3.2 Calibrate on curve

Using the optimum conditions stated above, a calibration curve for the

simple diffusion method is six own in Figaro 21. Sample solutions with

ammonia concentration in the range 0.1 to 1 ppm can be determined success¬

fully by this method. The correlation coefficient of this curve is 0.9915,

and the average root mean square error is 0.009 pom for that range.

Since it takes at least 13 ml solution for the normal 1 cm cell of

the spectrophotometer, and the ammonia absorbing solution is just 3 ml in

volume, so the method of using longer cell to lower the detection limit

cannot be used. If longer time for diffusion process is tolerable, so as

to condense the ammonia of larger amount of sample solution: into smaller

volume of absorbents, more diluted solution: can still be determined by this

method. For more concentrated ammonia solution, it should be diluted to

the required range before using the diffusion method, or using larger volume

of absorbent which can produce the same dilution effect.

59

Absorbance

0.8

0.6

0.4

0.2

0 0.2 0.4 0.6 0.8 1.0

NH3 conc., ppm

Fig. 21 : Calibration curve for simple diffusion method.

(5 ml sample soln.; 5 ml 0.01 M NaHSO4 as absorbent;

5 ml 5 N NaOH as alkali; 50°C for 4 hours)

60

3.3 Interference study of the simple dif f ws.ion method

The purpose of this simple diffusion method is to eliminate interference

of inorganic and organic compounds of direct Berth'.-lot reaction molted

From the result of interference study on this method it is found that this

method is successful. All the inorganic ions that interfered with the

direct Bertholot reaction method do no harm to the simple diffusion method.

Although some nitrogen containing organic compounds (e.g. urea, hydroxyamino,

some primary amines and anodes etc.) partially hvdrolv.,e in alkaline solution,

most of i hem are rare in natural water or polluted water vh5 ch are the main

objects of this method, and results show that these organic compounds shew

little interference to this method when their concentrations are below 1 ppm.

3.4 Conclusion

When sample .solutions show large interference by using' direct For the lot

reaction, the simple diffusion technique should be tried. Compared to other

extraction methods for ammonia (e.g. distillation method), the simple diffusion

method is much simpler. No special technique and apparatus are required.

Especially when it is combined with Northelot reaction method for ammonia

determination, it has the advantages of low detection limit, quite accurate

and little interference for general purposes.

This method is most suitable for the determination of ammonia content in

poD.luted water, sea water, and waste water. Ammonia in air can also be

detected after it is absorbed by a solution.

61

VT„ APPLICATION

The waste water of the Chinese University of Honp Kong is treated by a

sewage treatment plant before draining into sea. This is used to filtering

out the sewages and harmful materials, so as to avoid polluting the c-nviromen4.:

The ammonia content of the waste water, before and after treatment, was

never tested by the authority in the sewage treatment plan t of 0 U.H.K..

The methods described, in present work have been employed for the determination

of the ammonia contant.

1. Direct Nessler's method

Then Nessler reagent was added to the sample solutions (waste water

before and after sewage treatment), turbidity was immediately noted. The

colour of the product was different from the normal reddish brown colour,

and some white precipitates; probably alkali-earth hydroxides, occurred.

After standing for a few minutes, majority of the colloids settled down to

the bottom of the volumetric flasks. These particles showed large interferences

to the reaction. Therefore direct Nesslerization could not be used to these

ki nd of sa rrm 1es.

2. Direct Berthelol reaction method.

When completing agent and reactants were added to the sample solutions,

blue colour formed rapidly. However, the ammonia concentration in tin waste

water before sewage treatment was too high to be successfully determined. Ai

least a 100-fold dilution was required. By comparing the absorbmces with

the calibration curve, the ammonia concentration were found as in Table 8.

625.Shiple di-ffusion me fchod

Same as the direct Her the lot reaction .ethoa, the waste water before

sewage treatment was diluted 100 times before putting into the simple diffusion

flask, but the waste water after sewage treat me 3 if could bo used directly

After the absorption process, the absorbent solutions with the manoria

absorbed were tested by the Ber thelct reaction. .othod, eat the ammonia coi-

centration were determin. Trie results were cor;:;--red to that of direct

Berthelot reaction method in Table 8. It was found that the results by simple

diffusion method were slightly higher than that by direct Barthelot

reaction method. Tie difference might be arises from the interference of

inorganic ions in the waste water.

Table 8

The ammonia confent i 11 waste v;ater of C.U

before and after sewage treatment

Anmonia condent

before trea trnent

Ammonia cont ent

aftor treantnent

(ppm) (ppm)

Direct Berthelot

reaction method

Simple diffusion

method

38.540.70.080+0.005

41.0+0.3 0.520+0.055

A11 data are average of triplicate results with standard

deviations shown.

4. Conclusion

(68)The permissible level of ammonia has boon established as 0.5 mm.I

63

(i.e. 0.5 ppm) in waters designated as public water supplies. Aquatic life

exposed to levels of 1 mg/1 (1 ppm) may suffocate because of the significantly

reduced oxygen combining capacity of blood.

The ammonia content in the waste water of C.U.H.K. is quite high be-

fore the sewage treatment. The sewage treatment plant here is quite

successful in reducing the amount of ammonia in the waste water, ammonia

content is lowered by 98% when it comes out from the plant. Its concen-

tration is only 0.52 ppm. When it is further diluted after discharged into

sea, it is well below the permissible level of ammonia.

64

1. E.P. Odurri Fundamentals of Ecology, 2nd ed. V. b. Saunders Co.

Philadelphia and London 1959, pp. 30.

2• W.F. Ganon~? Review of Medical Physiology hid: ed. Red: of

Publication C' iiul ti} i e d ,i J ap a n 19 op% pp. 57 6.

3• N.H. Ferman Scott1 f; Standard Me thod of Che ideal Analysis,'' 6th ed

P. Van Nostrand Co. Inc? Princeton Ken .Tei sev Vol. 13962, oh.

4. J. N e s s 1 c r oh CP1• ne nt r. 27 N.F.I. 52 9 (1866).

5. AJ. Vogel. A text-Book of Quant at ive Inorganic Analysis 3rd ed;;

Longman 3 .o ncion 19oi pp. V83

6. J .P. Thompeon and GRP. Morrison Anal. Chen. 23 13 53 (195---)

7. G.L. Mil] er and E.K. Miller. Anal« Cheio. 20. 301(1938).

8. D.r. Boltz ed. Colorizetrie 33)eteriainat jon of Nonfat; Is Internedencc:

Publishers Inc. Ner York; 3-950 ppf35.

9. G- Fels, and R. Veatch Anal. Chen., 31 351(1559).

10. G. Jacobs Analyst 95; 3 70 (1970).

11. J .M. Prase and M .G. Mellon Aral. Chern.. 25 1188(1953).

12. Lidmil a Prochazkova Anal Ohom., 36.(865(1963).

13. V.W. Truesdale Analyst 96 583(1971).

14. D.N. Kramer and J.M. Sech jr3l_.._Che:Ti. 33 395(1972).

15. A. Levitzki A no 1. .,_B 1 oclpeiri 38 335(1970).

16. R.N. Hager Jr. D.R. Clarkson and J. Savory Anal_-L Ch en« 8-2 1838 (3-970).

17. H.J.M. Bowen Radio che;nr. radioanalvt. Lett. .3 107(1970).

18. G.I. Good fellow and H.M. Webber, Analyst, y.7 95(1972).

19. J.P. Riley Anal. Chirrw Acta 9- 575(1953).

65

20.

21.

22.

23.

24.

25.

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

42.

M.P. Berthelot Repertoire de Chimie Appliquee 288(l859).

S. Cotton. Bull. Sac., 21, 8(1876).

P. Thomas, Ibid, 11, 796(1912).

D.D. Van Slyke ana A. Killer, .JD Biol. Choia.., 102, 899(1933)

J .A. Baobell, A.Bfole. Biol C:i::», .156, 8570-988).

M.I. Kulenok, Gir 19229_i. jail5' 10, 1:5(1950).

W- G. Frankerbarg, et al., Anal. ..Ch em25, 17 28 (19.53)

A.Be Crowther and )b3. Large, Analyst, 81, 641.3.956).

B. Lubochinsky, and J .P. Za.lta, Bull. Chim. Bic.1 3.6, 1363(1958)«

R.H. Broun et cJ Arch. Bioohem. Biophys., 66, 301(3957).

L.T. Farn, «jr«; A n alCh a -.u, 35, 2179( j 963).

V.Z. Bohnstedt; Anali_Ghnni 163, 8.15(1958).

S« Solouay, and A. Sanucro, AnalChen., 27., 798(1955)-

U.Tv Bo] later 0 r J. Bus a nan and P.J. Tddvell., Anal- dlvaf., 33, 552(1963).

J-E. Haruood rad D.J. Huysor, bat... Res., 8, 501(1970).

V.T. Kaplan end V.G. Batsko Xzu. Akad» Nauk S_.3.3._R. Otd. Khlm. Mauk,

1673(1959).

P. Stegemann and B. Loescheke, Hoppe-Deylors Z- Physiol Cham. 329,

2 Id. (1962).

H. Well, R o i i,' -1| La. p» P' x j j. g 3 L77; L 182(1962)

D« B. Horn and CcR. Squire, 03.in. Chim. Acta, 3.7, 99(1967).

C.W. Gehrke, F.E. Kaiser and J.P. Uscary, Journal. qf the A.OnA.C.« 53.,

200(1968).

J.A. Tetlow and A.L. Wilson, Analyst, 89, 853(1968)

R.T- Emmet, Hav. Ship Roe. Develop. Cent• Rep.,. 2570(1968).

K.H. Mancy, Instrumental Analysis tor Water Pollution Control, Ann

Arbor Science Publishers, Inc., Ann Arbor, Michigan, 1970, pp.1.58.

66

43.

55.

A.L. Chancy and E.D Marbach; Clin. Chem.,, 130(1962),

V- Leithe; The Analysis ol Air Pollutants; Ann Arbor Science Publishers;

Ann Arbor; Michigan; 1971; pp. 172.

45.

56.

N. Tarugi and F. Lenci; Boll. Chim. Farm. JO; 907(1912).

L. N. Lap in Trudy Komissii Anal Khim. Akad. Nauk S.S.S.R.; Inst.

Geokhim. i Ana1. Khim.; 5, 7 7(1955).

57. H.S. Booth; ed.; Inorganic Synthesis. McGravr hill Book Co.. Ft j York.

193 9; pp. 59.

58. E.H. Kodd; Chemistry of Carbon Compounds;- Elsovier; Amsterdam V.ilJ B;

pp. 721.

59

50.

51.

52.

53

E.H. Rodd; Ibid; up718.

I. Weil and J-»C. Horris; !L«.__Am?. .Cbe.m. Soc..; 71; 3.665(1959)•

W.S. let calf; J. Chem. Soc.; 153(1952).

F.W.. Czech. R.Jc Fuch z and H.F« Antczak; Anal. Cheru, _;:3; 7;,5(19cl).

B .3. Newell J_- _Kar v_ Bio. Ass. B .K., 57 2 {1 (1967)•

J5. T ,L. Gilchrist and G «W. Recs Carbcnes, Nitrer.es; and Arynes-

Applatoss-Gentnry-Crofts Co. N.Y. 1969; PP-99.

55 Beilsteins Handbuch der Organischen Chemie Julius Springes.. Berlin,

1925; Band VIi; pp.619•

56. R.C. VJeast; ed.; Handbook of Chemistry and Physics, 53rd ed.; The

Chemical Rubber Co; Cleveland; Ohio; 19f2? pp. 1310'i.

57. B. Chapman, G«H Cooke and R. Whitehead; Pollut. Control; 66,

280(1967).

58. R. Th. Roskam and D. do Langen; Anal. Chime Acta; 30; 56(1965),

59» T.E. Weichselbaum and J G. Hagerty; Anal. Chem», 5l; 858(1969)«

60. J.B. Field; J. Coburn; J.L.- McCourt and W.A.E. McBryJ.e, Anal. Chim.

Acba,7b, 101(1975)

67

61. L. Solorzano, Limno. Oceanogr., 74, 799(1969).

62. M.W. Weatherburn, Anal. Chem., 39, 971(1967).

63. R.L. Searcy, N.M. Simms, J.A. Foreman and L.M. Bergquest, Clin.

Chim. Acta, 12, 170(1965).

64. A.L. Chaney and E.P. Marbash, Clin Chem., 8, 130(1962).

65. J.T. Wearne, Anal. Chem., 35, 327(1963).

66. E.D. Conway, "Microdiffusion Analysis and Volumetric Error,"

Crosby Lockwood and Son Ltd., London, 5th ed., 1962, pp 101.

67. E.D. Conway, Ibid, pp.102

68. H.F. Lund, ed.,, "Industrial Pollution Control Handbook," McGraw-Will,

New York, 1971, pp.4-30.