SPECTROPHOTOMETRIC DETER INA ~ION OF IRON IN ORES

WITH DIMETHYLGLYOXIME
by

DEAN OORE ROBERTSON

A THESIS
aubmitted to
OREGON S'l'A'l'E COLLEGE

1n partial fulfillment ot
the requirements tor the
<Iegree ot
STER OF SCIENCE

June 1952
IG?il0Er
Redacted for Privacy

effiEr rf [|r
Redacted for Privacy

Redacted for Privacy

€hrft*r of llLrnl SSrb fidttrr

Redacted for Privacy

3rraff$affi,h ffi"

nic srrl* li
w rteair frr IF.ntP
".:*
 '1'he writer wishes to acknowledge the valuable
aas1atanee ot Dr. J. P. Mehl1g 1n the expel"!mental
work and the organ1sat1.on or th1a tbes1a.
 '!'ABLE OF CONTENTS

Page
./Introduction l
vTheoret1oa1 Oona1derationa s
vExper1menta1 6

rApparatu and Solution& 6

~e Color Reaction 9

~Order ot Adding Reagents 9

~Intluenoe ot pH 9
Other Intluenc1ng Factors 10
Wave langth ot Maximum Absorption 11
v p1g. 1. Spectral Tranamlaaion Curves 12
lecular Extinction Coefficient 13
able I. Valuea ot Speo1t1o and
Jlolecular Ext1no.t1on Coett1cienta 15
. . . Pig. 2. Conformity to Beer'a r.w 16
l)etermination ot Iron in ONa 17
•
Table I~. Reaulta Obtained with
D1methylg1yoxime. 19
Table III. Duplicate Reaulta Obtained 20
Results 21
D1acuaaion 22
Summary
1 Bibliography
 S.PECTRQ·PHOTOMEi'RIC DETER !NATION OF IRON IN ORES
WITH DIME TffYLGLYOXI ~m

IN!fRODUCTION

Iron is most or·t en determined by oxidation and

reduction t1tr1metr1c methods . Electrode deposition
has been used and the colorimetric determination by
the use of the thiocyanate method 1s well known .
Other C·O lor1metr1c methods have been examined and
developed for ores by ehl1g and others (6, pp . 76•
79;· 7, pp . 52-S5; 8 , pp . 8 - ll) .
The present system employing d1methylglyox1me 1s

another case of applying a color.1 metric procedure to

the quantitative determination ot iron 1n ores . A
second aim ot parallel importance is to supplement
work by Mehl1g (5# pp . 4• 7J 6, pp . 76- 79) which supports
the view that macro constituents may be determined
quantitatively by colorimetric methods and the spectro­
photometer .
D1methylglyoxime 1a undoubtedly best known tor
the determination of' nickel , but it has tound use as a
reagent for other group eight elements , palladium and
p l atinum (18 , p.p . 7l- 73 ; 15, P. 40l) . Other elements
as copper , bismuth, cobalt and iron have been noted
 2

to react with this reagent (18, pp.71~'73}. It is the

last ease that is ot interest here. 1th i~on, a pink
to red color is developed which is often mistaken tor
the n1okel reaction.
This reaction of ferrous iron with d1methylgly•
oxime was tirst s~r1ously considered around the turn
of the last century (15, pp.40l-404; 16, pp.376-377)
when a visual eolor oomparlson procedure was developed.
:iore recently 1t bas been considerably refined and
extended to other d1oximes (11, pp.423-426J 10, PP•
1017-1020).
So far as can be ascertained the system has been
employed only tor determination ot trace or slightly
larger amounts of iron 1n materials. It is intended
here to show that 1t may also be used to determine
iron present as a major constituent in ores.
 3

'l'HEORETICAL CONSIDERATIONS

In order to rel te the colored solution to the

concentration a measure or the color ia necessary.

Absorption or radiant energy as light 1a the basis ot
the measurement . One ot the fundamental laws, Lambert's
law, states that each layer ot equal thickness or the
absorbing medium will absorb an equal traction ot the
radiant energy that traverses it . The decrease in
intensity tor traversing a length, 1, 1s then,

-~ • k'I where I 1s intensity and

dl
k' is a constant ot proportionality. It this is
integrated using the limits I 0 at 1 equals o, and I
at 1 equals 1, the result 1s,
ln Io/I • k'l or, I • I 0 e- k'l .
A second law with which the first is combined ia
Beer's law. This law expresses that the absorptive
capacity ot the colored medium 1a directly proportional
to the number or absorbing ent1t1es (concentration) .
Here the decrease in intensity with increased concentra­
tion may be formulated,
I • I 0 e-k''c where k'' is again
a constant ot proportionality and c, the concentration.
 4
Grouping into a new constant, k, the constants k',
k'', and the factor tor changing from natural logs to
log to the base ten, the two laws can be combined -.

losloio/I • kcl or, I •r 0 x 10-kcl

The factor, loglOio/I, is termed the extinction which
is an expression that wil~ be used later. The new k
is the spec1fic · extinct1on coefficient and indicates
the absorption due to a single molecule which is cons­
tant if' the law holds . ·
In order to put the measurement of absorption on
a workable basis 1t is necessary to allow for losses
ot light due to scattering and reflection at the two
interfaces of glass and air, tor instrumental scat­
tering and other error. Thia is eliminated by
measuring the intensity of a ray of monochromatic
light passed through the colored solution and also
the intensity or the ray through a colorless 0 blank•
of similar refractive index and contained in a well­
matched cell. This practice will give the rat1o I/I 0
Which 1s termed the transmittancy. If the basic law
is recalled it 1a seen that the absolute intensity is
not needed, but only the ratio is a1gn1f1cant.
For single colored constituents there are two general
procedures that may be followed to obtain concentration
 5

ot the desired constituent from spectrophotometric data .

In either case 1t 1a necessary to examine the color
.
system using known concentrations ot the substance
under 1riveat1gation. In the one case it Beer'a law !a
followed, i.e . it absorption increases proportionally
with the increase in concentration, the value ot k may
be calculated from the experimental values tor the re ­
maining terms ot the basic law equation . This k 1n
turn then 1a used for calculation of the per cent
composition ot the substance 1n the sample .
In the other case , where Beer's law tails, a oal1•
bration curve may be drawn from the experimental
values on a aeries of solutions of known oonoentra­
tion. !be concentration ot the unknown sample !a
then read ott from the curve. Examples in the lit­
erature may be cited (5, pp . 47J a, pp~S-11) ~

A fuller treatment of the theory and application

ot color systems is given by Mellon (9, pp . SS- 101).
 6

EXPERIMENTAL

APPARATUS AND SOLUTIONS

Cenoo-Sheard Sgectrophotelometer. The instrument used

to obtain the ratio ot I/I 0 was the Cenco-Sheard Spectro­
photelometer.. The entrance and ex1 t ali ts were used at
set values and the intensity of the 1no1dent light was
regu.lated over the range of wave lengths by means ot an
iris diaphra.m between the light source and the entrance
alit~ However , the d1aphram would not accomodate the
entire range. The exit alit w s set at 20 liJl tor the
ranges 400 to 440 and 680 to 700 mp, a 10 ~ alit tor
the r-anges 440 to 480 and 640 to 680 Ilfl~ and the 5 ~

slit was used for the middle range of 480 to 640 ~·

The iris was regulated with the •blank" in the light

path so that the galvanometer read 100. The colored
solution was then exchanged for the blank by means ot
a sliding carriage. Since transmittancy is the ratio
or the intensity of light passing through the colored
solution to the intensity ot that passing through the
blank, tbe reading of the colored solution as registered
on the galvanometer seale is the per cent transmittancy
directly.
The oells were l•cm. 1n length w1th pyrex windows.
 7

.e aame cella were u•e4 tor all a ure enta. one

uaed exclua·1velJ tor the blank and the other onl7 tor
the 40lored aolutlon.
Beckman pJ! ter. All pH aaurement were made with
a Beckman el R•2 pH ter.
Dl ethzlglyoxlme. A aaturated aolutlon ot th reagent
ln 95. ethyl alcohol waa u.aed. 1'b1a waa detenllned to
be approx1matel7 a 2.39 per cent aolutlon at 26 degreea
centigrade. 1.'he aolu.t1on e:xb1b1ted no apparent break­
down w1 th tl •
Standard Iron Solution. A a ple ~ we1gb1ng 0.2250
gram. ot an iron ore, oontaln1ng 14.44 per cent !ron
aa 4•term1ne4 b7 the d1cb~mate method (4, p.242) we
treated With concentrated b,-drochlo~lc ac1d 1n a 250­
m1ll111ter beaker and warmed on a bot plate at low
beat until aolutlon waa complete or only a amall
white a111ceoua realdue remained.
!he cooled aolutton waa tranaterred to a lGOO-ml.
vol\1Jiletr1c tlaak and mad• up to volume at a regulated
temperatUN. Each 1llll1ter ot the rea:ulting aolutlon
contained 0.1 mg. ot iron. Also eaoh milliliter taken
and made up to 100 ml. gave a aolut1on wboae coneentra•
t1on waa l P•P• of iron.
H7drox;rlam1ne Htdrochlorlde. An aqueoua aolut1ou con•

ta1n1ng 10 grama per 100 mdll111tera.

 8
Ammonium Chloride~~· An aqueous solution containing 200
g;rams ot the salt per 11t.e r.
Ammonium Hydroxide . A commercial 15 [ solution.

I
 9

THE COLOR REACTION

When ferrous 1ron ia mixed with dimethylglyox1me

and ammonium hydroxide added a pink- to deep red•colored
solution develops. The greater the concentration of
iron the deeper the red color.
If dimethylglyoxime is designated as D2H2 and a
· molecule of ammonia as A, then the complex could be
written as (D2H)2A2Pe. Although this is the usual
composition aometlmes three molecules of annnonle. are
tound aasoe1e.ted with the complex (15. pp.40l-402).
Order ~ Adding Reagents . A study or the effect ot
the order of addition ot reagents on the color reaction
revealed that the order 1s rather critical. The 1ron
-should be reduced 1n an acid solution to attain tull
color production in a short period of t1me. The d1­
methylglyox1me must be present with the reduced 1ron
when the mixture 1a made basic and all the reagents
cannot be mixed and the !ron added at the last. Some
other orders than the one used develop maximum oolor
after stand.ing tor a time, but tor color development
1n ten to fifteen minutes the present procedure should
be followed.
Intluence 9l. ~ Since hydrogen ion 1s involved 1n the
color reaction it might be expected that the pH value
 .

10
would have an influence on the color of the solution
and that is the ease. The color increases as the pH
value increases. Following the recommendation of Baker,
(11, p.42:S) the pH was controlled at a value of 9.4
to 9.5. This was accomplished by th~ use of an
ammonium chloride-ammonium hydroxide butter solution.
Checks were made with the pH meter from time to time
throughout the experimental work in order to verify the
correct value.
After obtaining a correct absorption value it was
tound that control ot the pH value between 9.2 and 9.8
was suf'1''1c1ent. The extremes ot these values gave a
change in transmittancy of the order or 0.1 unit which
is within the range or acceptable error.
Other Intluencing. Factora. The effect ot diverse ions
on the system has been studied and reported (11, pp.
423-425; 10, p.l019). Some twenty-nine ions are
covered in theae articles along with their maximum
ratio to iron by weight that can be tolerated.
The color fades with time and the stability was
determined by making transmittancy measurements at
convenient time intervals. It was round that the more
concentrated iron solution was the least stable and in
regular order the stability increased with the least
'
concentrated solution containing 0.5 p.p.m. iron being
the moat stable.
 ll
The stability ia also lntluenced by the reducing
agent. The tadlng of t he color is due to the reoxida•
tion of the ferrous to ferric 1ron ln ammoniacal solu­
tion by dissolved and atmospheric oxygen (12, p.lll3).
Sodium dithlonate bas been used as a reducing agent
and is said to have a stabilizing effect on the colored
complex (1, p.71; 10; P•lOlS).
Hydroxylamine hydroch loride has been used success­
tully by various workers (10, pp.l017•1020; 11, PP•
423-425}. All color solutions in this work were made
up with recently boiled and cooled, distilled water and
were not exposed to the air any more than necessary.
For the best results the transmittancy readings should
be made w1th1n 24 hours ot the time the colored solu­
tion is made up to volume •
..!!!. Length .2!, Maximum Absorption. The wave length of
maximum light absorption was determined by taking the
absorption measurement ot five solutions with concentra­
tions or o.s, 1.0, 2.0, 3.0 and 4.0 p.p.m. ot iron.
The standard iron solutions were made up by
measuring suitable al1quots from the standard iron
solution into a 100-ml. volumetric flask; carrying
out the color reaction and diluting up to volume at
thirty degrees centigrade with boiled and cooled; dla•
tilled water. e~actly as described 1n the procedure for
the determination ot iron (p.l7).
 12
-·· '-- ., __ !__.. .
'
'
I

'
I
0-t­
l­
I
t
i
l • I ~
I ! ~
,. .._ L--·-~- _j -·---~ --~---+-
' · i
.
I I

a 1 '
~ -­ .. I
. '
'
'
··----~ ---~
i
-z --­
I
1

1
--·. ·r --­
1
I
....
i
 1

.eot al tr 1 ion o~ve ere o t 1ned u

ho 1n P1 • 1~ Th 1 ngth ot 525 waa th&

a the polnt u 4 tor tbe d t rm1nat1on ot the ores.

lecula~ Extinction Co ttic!ent. 1 • l e al o
co tro t tr n 1 ttancy t 525 r t t1ve
olution oont 1n1 .0 P•P• •
.
ot iron. ntrance alit wid h a ti% at 1.5
and the ex1t lit width at •
It wa re91ou 11 d te 4 that 1.t Beer• a law
1 obeye4 'h ap o1t1o xt1no 1on tt1elent aa
det 1'm1n4 t r e 1 n th 1 a ctmat nt !h
leoular xt1nct1 n c ttl len ta r 1 th •pee1t1c
-coett1 len lt1 ~1 4 T tne le 1 t ot th
con tltue t to b et r 1ne • It t ou.ld be a
o natant.
n the xt1nct1on value were calculated and
plotted a 1 t the pect1v conoentratl n a
tra1ght line aulted. ee Pig. 2. 'l'hia 1nd1c ted
that Beer' law a obeyed and the caloula ed value ot
the l~cular ext1nct1o co tt1o1ent. e# a c~natant.

The caloul t1on 1 de rro the er- ort equ tion

aa tollowa.
 14

loglOio/I • kcl aolv1ng tor k,

k • loglOl:o(I and,
C X l

e • lilk lrhere 11 ia 55.95 or the

atomic n1ght ot iron and tbe other aymbola have the1r
prev1oualy stated aign1t1.cance.. 'lbe average value or
e waa calculated to be 9498. See table I.

.
 15

TABLE I

VALUES OF SPECIFIC AND OLECULAR

EXTINCTION COEFFICIENTS

Cone. ot Iron Spec1t1c Extinction Mol . Extinction

p .p.m. Coeft1c1ent Coett1c1ent

0.5 171 .3 9567

1.0 170.1 9500
2.0 -170.5 9522
3.0 169.5 9467
4.0 168.9 9433

Average 170.1 9498

 .l.f7

DETEUINA·TION OF IRO:tl IN' ORES

About ().•4 gam ot aam.ple wae eecw:-ate~y welgbe4

and transte:rred : o ·a 250-ml. beaker~ ~e ore was
then ~ted wtth 26 ml. :et c~~entrate-d hydr-~chlor14

aoid,: eoverad wt th a wat.eh, t;Jlaaa, and lett ~n a ho"

plate at low heat unt!l :itOlut1on was complete or ~nlJ

l1gbt colored a111oeous .r esidue rema1n•d• The

coo-led solution weuJ. t'llltered Into a lOOO.ml. Yolumetr1c
tlaak and the rea 1due. was wasb..ed w1 th 300 to 400 ml.
of' approximately o.OlM hydroehlor1c .acid solution.
The t1l.trate waa made up to volume ith d1st1lled

water at 30 d~geea oant1grad,e,. and thoroughl.1 shaken.

With a. llloroburette 1 ml. ot the ir-on solution
w-aa carefully mea.a ured S.ato a lOG-.ml. volumetric tlaak
and d1luteil wt th bo1.l e4 and o.ooled, cUatilled water.•
TlHtn 1n 1)rder- lfe·:P& add.e d 2 ml. et the hydroxylamine

bydr-ocblortde aolut. 1on~ 22-.-50 ml. of the ammonium

chloride solution., and 2 ml. ot the d1methylglyo.x1me
reagent. After shaking, 10 ml. or the 1511 ammonium
hydroxide were lntroauoed, the mixture was made up to
volume w1th distilled water at 30 degrees, and
thoroughly shsken.
 18
The transmittancy was measured at 525 and the
percentage ot 1ron was calculated by use ot the
molecular extinction coef!1oient. The following
example illustrates the method ot calculation. ·
Experimental Data: weight ot ore sample; 0.4000 gram;
transmittancy of the oo).ored solution at 525 ~~ 55.4%J
length ot cell, 1.0 em., e; 9498 . Solving the Beer­
Lambert equation tor o gives;

c .• loglOIQ/I moles ot iron per liter ·

e X l
Since Io/I • 1/~ where T ~· the per
cent transmittancy,
o • loglol/T x at. wt. Pe grams P8 per liter
eX l
Since the final solution contained 1n 100
ml. o•.OOlth ot the or1g1nal solution,
f.Fe • los10l/T x at. wt. Pe x 1000 .x 100
e x 1 x wt. ot sample x 10
%Fe • los10l/0.554 x 55.85 x 100 x 100
·§49§ X leO X . 0.4000
'/le • 37.70
 1

Pe~nta •
ro

1 3?. 37.
•
-o.
•
o.ovv
+0.01 o.OZ?
. ~ M• -o.
56. 68.06 + 06 .on
' 57. til. +o. .06SJ
.ao 52.1? 0.03 o.o
se.ctt -o.CH o.ov
57.i0 &'1. e>.oo o.ooo
9 1. l.SS +O.OS o.o
10 9. 4,9.54 -o.a 0.101
 20

·TJ..BLE III

DUPLICATE RESULTS OBTAINED

sample
lfo.;..
&II f6iii - Difference Penentage
Dev!.._t1on

-~ ~ f. ~
1 37.• '70 3,••, 0.23 0.61
2· 36.88 36.81 0.07 0.02
3 53.9'1 54.07 o•.10 0.02
4 56.11 55.96 O.. l5 0 .• 27
5 57.66 57.66 o.oo o•.oo
6 52.22 52.ll 0-.ll 0.21
7 52.8. 52.'7:5· 0.11 0.21
8 5'7.~ 5'1.96 0.13 0.22
9 51.6$ 51.• 4$ 0.15 o.st
10 49.. 54 49... M o.oo o.oo
Av•rage 0.11 0.19
 21

RESULTS

The results obtained tor ten ores are shown in

Table II together with the values given by the di­
chromate t1tr1metrio method. The greatest difference
between the two methods was 0.05 per cent with an
average ditterence ot 0.03 per cent. The percentage
error varied trom - 0.101 to +0.071 per cent with an
average ot -0.01 per cent. The d1tterence between
duplicate results is shown 1n Table III. Best results
were obtained by taking the average ot a large number
ot transmittancy measurements .
 22

DISCUSSION

The result$ ot the analysis or the ten ores would

indicate that this spectrophotometric method can be
used tu measure 1ron aa a. maaroeonst1tuant. This 1a
shown ·b y the eloae ag~eement with th.e value.a obta1nttd
b7 the recognized t1tr1metr1a method {4, p .. 242).
1b.e method has been used tor determination of
izson in wate.r (1 1 p.71} by a less exact method ot
visual comparison 1n Nessler tubes. ~· present
method haa the advantage ot not requiring a series ot
standards tor compar1 on.
.
"The r&'act1on bas been used to ,d etermine ferrous
ln the presence or fe:r-r1o !ron by use or tartaric
ac1d (17, p.l5).
lfbe method has the advantages ot (1) tew 1nter­
terencea• expeo1ally very few in the ease ot ores,
(2). nearly immediate formation ot color, {3) con­
:rormity to Beer•a law in a workable range of con•
centrat1on. (4) easily obtainable and 1naxpens1ve
reagents and (5) 1ndependanea of excess of the reagent.
 23

A spectrophotometric method tor 1ron 1n ores

has been cleveloped, wh1eb depends upon the pink- t.o
red-eol&.red, aolubltll complex tormed b7 ferrous iron
with dime·thylglyox1me 1n an ·ammonia-cal s.olut1oA~

'l'he apte·m was round to conform to Be.e~ta law

tor the conccentrat1on rang& ot o. 5 to 4 P•P·•• ot
1ron. 'ftle peak or the absorption band, 526 . , waa
uaed tor the dete:ron1nat1on of the molecular ext1nc.t1on
coef't1c1ent, which relate,s trans.m ittane,- and con­
centration or the l:ron.
The reault.a demonstrated the accuracy ot the
method by tavorable compar1aon w1th the raaults ot
the dichromate t1tr1metr1c metbod.
!he method 1a eaa!l.7 .operat.ed ana can be used
tor ores conta1n1ng up. to at le-ast 58 per cent iron.
It has aeveral advantageou-S aspects w1 th -very tew
interfering diver·ae ions where iron ores are concerned.
 24

BIBLIOORA Y

l• D&unce,-, L.. H• Color1Mt~1o deteNS.nat1oD or a 11

quantltlea of lron 1n water b7 •ana of
dimeth7lglyox1me. ~· analyat 74:?1. 19&9.
2-. Kraua • E. J. Ueber d1e Bealtt1on dea Peno-Iona •1t
D1- tnJl&l7oz1•• Ze1taonr1tt tUr anal7t1&ohe
Chellll• "1ltl89•190. 19-27.
3. Lanae. orbert Molpb• (ed.)• RaadbOok of
obe111at17'• 6th e4. 8anduaq. Ohto.
Handbook publlaber•~ 19•6• l'167p-.
•· in. •dwa-d o. Quant1tat1~• anal,.a1a. •tb ed.
• ·York. McGraw-Bill. 1932. 62~p.

s. hl1g, Jo•eph Parke. Speetroplloto trio deter-­

.S.natlon of a.1okel ln ateel lth n1a.
Ohelltat aoal7at 3' a<l-'1. 1945.
&. ~---- and Doroth7 llsabetb ~at. Appl1oat1on
ot aa11o7laldox1me to the apeotropboto tr1c
4etel"JJ1ruat1on ot 1Hn 1n ore•• Cheld.at
aul,.at as t'76-7i. 1949.
'7. --and urtce J. Sbepbn-4, Jr. Spectropboto­
•tr1o det~t1on ot iron 1n orea w1th
2•2'-b1p)71d71• Chemat analyat $6t52•65•
1N'7.
a. - - - and. urlce 1. Sb.epruil'd, Jr. Appl1oat1on ot
tb1o~l,-ool1cac14 to the apectro»botometrlo
deteratnat1on ot !.ron ln ore•. Cbe at aual,.at
~5a8-ll. l~U.

. ~. llon. 1Y1n Gu7. (e4.). Ana1Jt1oal abaorpt1on

apeotroaoop,-. • York. l obn 11e7 and Sou.
1050. 618p.
 25
10. - - - - ana llargaret Gr1i't1ng• Colorimetric Giet.e r­
m1nat1on ot iron: with various d1ox1mes.
Industrial and engineering chemistry
(analytical e~1t1on) l9tl017•1020~ 1947.
11. Rodden. Clement James. (•4• )+ Aualyt1cal chemistry
ot the m.anb.attan p!-ojeot.. .~:~w York. McGraw­
Hill• ltSO. Y48p.
12. Slaw1k. P. DlmethlJglyox1me aa a aena1t1ve
reagent for ferrous salts. Ohemikar...
ZeltU%11 ~6t54. 1910.
1.3 . Snell~ Poater De6. Colorimetric analysis. Bow
York~ D.Van Xoatrand, 1921~ 150p.
14. ··-· -- and Cornell& T. Snell. Colorimetri·e metboCla
or analra:ts.V-ol. 2. 3d ed. Hew York•
D. Van Nostrand~ 1949.. 950p.
15.. Tachugaefr, L. and B. Orelkin. Ueber efne empt1nd­
l1che E1senreak·t 1on and ueber eine Methode ~ur
kolor1metrisohen E1senbeat1mmuilg. Zeitsohritt
der anorgan1aoh• Cbemie 89&401-404. 1914.
16. ----- and B... Orelkin.- lle.ue Reakt1on und kolori•
me-triac-he Best1mmung -des Eisens. Ze1 tschritt
tlir analy-t1s,cbe Cb.emie 54"376•3'17.. 1915.
17. Von Stein. Paul. Rap1c! color-1metr1c detertt11nat1on
or 1r(Jn. Chem13t analys~ 34:15. · 1945.
lS. Willard 1 I!obl\:rt fi~ a nd Rarvey Diehl. Advaneed
quantitative aualya1s. .3d ed. Hew York.
p.van Nostrand, 1946. 4i7p.
19. Yoe, Jolln I!Qwe.. fhQtometrie chemical analysis.
Bew Yorlt 1 John Wiley and Sons, Vol-. 1·
1928. 77Jp.