~&m, Vol 22, pp 98%993 Pcrgamon Pres$ 1975 Pmted m Great Britain

IODOMETRIC DETERMINATION OF ASCORBIC ACID BY

CONTROLLED POTENTIAL COULOMETRY

RONALDKARLSSON
Department of Analytical Chemistry, Chemical Center, Umverslty of Lund, S-220 07 Lund 7, Sweden

(Recewd 1 October 1974 Accepted 2 February 1975)

Summary-An xodometrx method for the determination of ascorbic ac:d has been devlsed The method
IS based on prevzously developed coulometrlc mstrumentation The stab&y of &fferent ascotblc acid
solutions has been stu&ed and the best condltlons have been estabhshed Ascorbic acrd has been
determmed m different kmds of samples but with the mam interest on pharmaceutical preparations
Spew1 regard has also been paid to the other constrtuents m stich samples, with respect to possible
mterferences The error of the coulometrlc method IS about 0 1% and the time of an analysis IS m .
the range 2-6 nun

A large number of methods are proposed m the cur- high accuracy IS obtained The esttmatton of ascorbtc
rent literature for the determmation of ascorbic acid acid by titration with iodme is m many countries an
(Vitamin C). Most are based on modtfications of the official method” for many pharmaeutical prep-
method proposed by Trllmans et al ,I where the ascor- arations. Rather serious drawbacks are that the
hc acrd is allowed to react wtth ~~chloroph~ol- trtranons with &me are restrrcted to use m relatively
mdophenol m a slightly acrd solutton. Thus procedure strongly acrd solunons and the todme reagent needs
1s the official method m several countries for the frequent standardizanon
determmation of ascorbic acid m fruit and vegetable In our laboratory we have already developed very
~uices~ and in pharmaceuttcal products.3 Although tt accurate analytical methods using the todrde-todme
1s an official method, it is not apphcable to many system and controlled-potential coulometry,12-14 and
p~r~u~~ preparations owmg to ln~rfering an electroanal~~ approach to the deter~ation of
ions generally present m such products One of the ascorbic actd with rodme as an mtermedrate seemed
more widely used colortmetrrc methods was devel- mterestmg The use of controlled-potential coulo-
oped by Roe and Kuether,4 who oxldrzed ascorbic metry for the determination of ascorbic acid has been
acid to dehydroascorbic acid and let the latter react proposed earlier by Santhanam and Krlshnan,15 who
with 2,4-dmitrophenylhydrazme to give a red colour, oxidized ascorbic acid directly at a platmum anode
which was measured ~~torne~i~y However, some with a potential of 1090 V us, SCF! The high oxidatron
other substances, such as 2,3-dtketogulomc acid, react potentd used made the method unselectrve for ascor-
similarly. S&mall and co-workers5+6 reacted ascorbrc bc acid The times of analysts were up to one hour
and with dtazotlzed 4-methoxy-2-mtroanilme 111acid and the error m the range 15-100 mg was not better
medium and measured the intense blue colour devel- than +07 mg
oped after addition of alkah Among other authors In the coulometrrc method presented m thrs paper,
using diazotrzed compounds can be mentioned Weeks based on the Connolly-pound apparatus devel-
and Deutsch,’ who used dtazottzed p-mtroamlme, oped earher, todme m a shght excess over the amount
and Hashmt et al ,* who used dtazottzed p-ammoben- of ascorbtc acid to be determined 1s generated at
zoic actd The latter method was spectally worked + 300 mV us SCE and after addmon of the sample
out for pharmaceutical products the remamder of the mdme 1s reduced at OmV
In general, the colortmetrrc methods suggested for us SCE and the result is directly obtained on an elec-
the assay of ascorbic acnd are rather time-consummg tromc integrator The tunes of analysis he m the range
and many of them suffer from lack of colour ,stab&ty 2-6 mm and the error IS about +0-l%
and are apphcable only m rather narrow ranges of
concentration Together with the method of Tillmans,
the methods tnvolvmg lodmey have received the EXPERIMENTAL
greatest attention. In the latter methods the followmg
reaction between ascorbic acid and :odme is exploited Apparms
The ~n~~ied-po~ntl~ coufometrlc cvcmt and the
C&H,O, f I, --+ C,H,O, electroivsls ceil have been described m detad earlier 13*‘6
+ WC + 21-
The e&trolyte m the’counter-electrode compartment and
Starch is generally used as mdicator Accordmg to bridge compartment of the electrolysis cell was 2&f sodmm
sulphate and m the workmg electrode compartment 1M
Stevens” the use of iodme has many advantages over sodium snlphate/lM sodium Iodide The pH m the work-
other methods proposed The analysis IS rapid and mg-electrode compartment was adlusted with acetIc acid
989
990 RONALDKARLS~ON

Reagents Table 1 Recovery of ascorbic acid, with different excesses

of iodine added
ExtractIon solutwn An 8% solution of glacial acetlc acid
was prepared for extraction of ascorbic acid from the sam-
Amount of ascorbic acid,
ples to be analysed Amount of lodme pmole
Standard asc&lnc acid solutions Prepared from pro ana- generated,
lvsr auahtv ascorbic acid (E Merck. Darmstadt) The solu- pmole added found
ilo& were prepared lmiedlately ‘before use klther with
doubly-distilled and oxygen-free water or with an oxygen- 400 201 207
free solution of 8% acetic acid 300 201 207
2 50 201 201
Preparation of samples 2 15 207 207
(a) Samples from the standard ascorbic acid solutions 2 10 201 204
were transferred directly to the cell with an Agla 2 10* 207 207
micrometer syrmge
(b) Samples from Juices, pharmaceutical solutions and * Allowed to stand for 15 sec. before reduction
other hqulds were either transferred &rectly to the cell
or diluted with the acetic acid extraction solution m care-
acid. However, a 0 1-O 5% correction must be applied
fully calibrated standard flasks Homogeneous and not too
vwous samples were generally analysed directly wlthout when the excess of lodme 1s not reduced lmmedlately
dilution, to avoid air-contammatlon This 1s due to small losses of lodme caused by adsorp-
(c) Samples from solid products, such as tablets, were tion on the walls and electrodes, by transport through
first powdered m a mortar, then shaken with the acetic the filter and by volatlhzatlon. Figure 1 shows the
acid extraction solution, filtered, transferred to a standard
flask and made up to the mark results from a series of measurements when 5 00
wale of iodine were generated m each determmahon
Procedure and 1, 2, 3, 4, 5 and’ 10 mm were allowed to elapse
The compartments of the electrolysis cell were filled with before the start of reduction The measurements were
the appropriate electrolytes The solution m the working repeated three times and the results agreed wlthm a
compartment was contmuously deaerated with nitrogen at
few per cent The linearity 1s rather good and the
a flow-rate of l&20 ml/mm Any lodme present m the
compartment was reduced at + 0 mV us SCE until a stable losses are very small, which facilitates the corrections
residual current was reached, generally not exceeding 5 It also appears that the total time of an analysis m-
PA Then iodine m excess relative to the amount of ascor- creases by more than the waiting period. Thn 1s due
bic acrd to be determmed was generated at 300 mV us to the fact that some of the adsorbed lodme slowly
SCE and the sample was added Then the lodme not con-
sumed m the reaction with ascorbic acid was reduced until goes back mto solution and then can be reduced,
the earlier residual current value was reached resulting m a prolonged electrolysis time The time
for a complete analysis was measured from the start
of oxldatlon unhl a residual current of 5 fi was
RESULTS AND DISCUSSION
reached
The redox potential values for a solution contam- Solutions of ascorbic acid suffer from a lack of sta-
mg ascorbic acid and dehydroascorblc acid have been blhty, ’ * and m earlier mvestlgatlons it was dficult
calculated by many authors17 and there are rather to monitor this mstablllty and control it during a
large discrepancies m the hterature Accordmg to Rao longer period, owing to the need for extensive stan-
and Raols the potential varies from -0012 to 0 326 dardlzatlon of the titration agents and cahbratton
V with variation of pH from 8 67 to 105, and our solutions, especially with titration methods employing
own determmatlons give a value of about 025 V at lodme In the present lodometrlc mvestlgatlon, how-
pH 3 This high reducing power makes ascorbic acid ever, this problem 1s overcome by the in szfu gener-
suitable for reduction of a weak oxldlzmg agent, such ation and back-reduction procedures used for the
as iodine The redox potential for a solution contam- Iodine
mg the lodme-iodide couple 1s about 0 54 V, calcu-
lated for the reaction
I; + 2e+31- 60 -
The reaction between ascorbic acid and lodme 1s
practically instantaneous and only a very slight excess s
of iodine needs to be generated, the remainder can H 40-
0
be reduced at once In Table 1 some of the results
of these measurements are collected As it takes only E

a few seconds to generate the amount of lodme 20 -

required, an excess of about 25% 1s generally pro-
duced, avoldmg the risk of mcomplete reaction
When cloudy or viscous samples are analysed, a I
00
relatively large excess of lodme 1s generated and up 0 4 8 12 16

to 10 mm are allowed to pass before the reduction m,”

1s started This will ehmmate any errors caused by Fig 1 Loss of lodme from the electrolysis compartment of
decreased reaction rate between lodme and ascorbic the cell as a function of the total time of analysis
 Determmatlon of ascorbic acid 991

Table 3 Determmatlon of ascorbic acid m some citrus

fruit Juices
100 . ..I .^._.._ -.__.___._.._._.
I.... .-.--,
Mm amounts Added amounts
; =I of ascorbic of ascorbic Found e~o,‘loOrnl
Kmd of aad’ aad
E -3
-1
,UlCL wq II10Ull mg/lW ml filtered unhlarLd
e! 090 p:
1
Orange 32 360 361
2 Orange
II 35 442 440
0 35 409
h Orange III 35 197 610
35 395 8@2 810
090 Grapefrlut 50 554 552
-.
25 311
Lemon 2s 39 5 709

* Accordmg to the manufacturers

hours

Fig 2 Stablhty of ascorbic acid m different aqueous solu-of standard solutions of ascorbic acid was prepared.
tions (1) 0 1M ascorbic acid m 0 05M oxahc acid (2) 0 1M Table 2 gives a schematic representation of the
ascorbic acid m 005M sulphurlc acid (3) 0 1M ascorbic
results All samples were added with an Agla
acid (4) OOlM ascorbic acid (5) 0 1M ascorbic acid m an
micrometer syrmge and were m the range 0 OlO-0.500
oxygen-free solution (6) 0 1M ascorbic acid m 0 05M oxalic
acld/@OSM sulphurlc aad, oxygen-free ml The workmg range can be expanded m both dir-
ecbons but at the price of increased tune of analysis
Five dtfferent ascorbic acid soluttons were prepared or lower accuracy, as the case may be
and transferred to 25-n-J standard flasks, two flasks An lodometrlc determination of ascorbic acid m
of each solution being made up All solutions were citrus fruit JUKXS has been shown to be the most reh-
prepared with doubly-dIstIlled water but, except for able method, since these contam, according to Tauber
one of the solutions, which was deaerated with and Klemer,” no substance which will interfere m
nitrogen, no precautions for excluding air were taken such an analysis Hence It seemed attractive to apply
in the preparation During the experiment all the the present method to some commercially available
solutions were allowed to stand oti the laboratory Juices Samples (0 l-l 0 ml) were taken both duectly
bench, exposed to light Samples were taken from from the packs and after filtration of the Juices The
each flask at intervals of l&15 hr The results are results agreed well, as can be seen from Table 3 The
collected m Fig 2 The curves are based on a least- blink values and the tnnes of analysis tended to m-
squares treatment of the values obtamed It 1s clearly crease contmually when the samples were unfiltered
shown (curves l-5) that the deaerated solution (No The cell solution must then be changed after a certam
5) kept its stability best and that the deterioration number of determmatlons if uncertain results are to
1s also much dmunlshed by addition of sulphurlc acid be avolded
(No 2) If possible, the JUNXS were not diluted, m order
The use of oxalic acid, which 1s often recom- to avold exposure to an The results had an average
mendedlg for protectmg ascorbic acid agamst 0x1- devlatlon of about 0 1% from the mean of three
datlon, also shows a slight improvement of stabrhty values
(No 1) m comparison with the untreated solutions The determination of ascorbic acid m pharmaceutl-
(Nos 3 and 4) On the whole the results show good cal preparations has without dbubt received most
agreement with the results obtained m wrmlar exper- attention Although there are several official methods
iments by Rao and Rao l8 The conclusions are that available, most of them are not applicable to many
oxygen-free soluttons are a necessity and that the pharmaceutical products, owmg to presence of dlffer-
solutions should be acidic and concentrated Figure ent kinds of Interfering substances Therefore we con-
2 also shows a curve (6) for the stablhty of an ascorbic sldered it of interest to test our method on some phar-
acid solution prepared 111accordance with these con- maceutical products, and chose a group of common
cluslons, fully confrmmg their validity multivitamin preparations
To test the coulometrrc method with respect to Table 4 gives the values for the determination of
time of analysis, working range and accuracy, a series ascorbic acid m SIXdifferent preparations Except for
Table 2 Determmatlon of ascorbic acid m pure standard the preparation “ABCDm” which 1s a syrup and was
solutions analysed directly, the other preparatrons were treated
according to method (c) (see Preparutlon of samples)
Added amount of Approx times We always used at least $10 tablets of each sample
ascorbic acid, of analyses,*
preparation m order to obtain relatively good sam-
pmole Error, % mm
pling as mdlvldual tablets vary 111weight by several
5-50 01-005 56 per cent
OS5 02-01 45 From Table 4 It can be seen that there 1s good
01-05 03-02 354 agreement between the results obtained by the official
0 01-O 1 2-o 3 2G3
methods of analysis2 lz and by our method No
* The times may vary somewhat depending on the excess notable Interferences were observed, but some mvestl-
of lodme generated gatlon was made of the best method for analysis of
992 RONALDKARLSWN

Table 4 Determmatlon of ascorbic acid m some pharma-

ceutlcal vltamm preparations
600 I’ ’ ’
Numberof Ascorbic acld found, n&abler
tablets
preparedfor by official by present
Type’ each detn methods method

IDO-CT”’ 5 516 522 5

BaSCOPkX’2 5 301
Mmorplex N”’ IO 43 3
Multlplex’4’ 5 64 I
Multiplex Comp N”’ 5 624
ABCD&’ syrup 513t
All preparations from Ferrosan AB, Sweden
*
j-mg/lOO ml
(1) Vitamin C
(2) Vltarnms B(B,BZBs), C. Ca-panthothenate and mco-
tmamlde
(3) Vltamms A, B(B,B,B,), C and D, Ca-panthothenate
and mcotmarmde
(4) Vitamins A, B(B,B,B,), C, D and E, Ca-pantho-
thenate and mcotmaxmde
(5) Vltamms A, B(B1B2Bs),C, D and E, Ca-pantho-
thenate, mcotmaxmde and Fe’+ (ferrofumarate)
(6) Vltamms A, B(B,B,), C, D and mcotmarmde
“ID0-C ” The only active ingredient 111this product
1s the ascorbic acid, but this 1s present m such a large
amount as about 500 mg m each tablet The samples
must generally be diluted m analysis of these tablets,
whichever method 1s to be used Because of the sensl-
tlvlty of ascorbic acid solutions to oxygen, erroneous
results can be obtained, owing to oxygen m the dllu-
tlon medium. The use of acetlc acid solution for dllu-
tlon preserves the ascorbic acid concentration some-
what, but according to Pontmgz3 a loss of more than
30% m concentration 1s observed within 24 hr In
an oxahc acid solution, however, the concentration
decreased by only 2 8% during the same time
We made three different sample preparations from
IDO-C, whrch together with the results obtained are
shown m Table 5 The effect of oxygen 1s again con-
firmed No effect of the oxahc acid was observed,
because of the short time allowed to elapse before
the analysis was done
Much effort has been made m the pharmaceutical
industry to prepare stable ascorbic acid syrups, and
a comprehensive review has been given by Hashml ‘*
We decided to study the stability of the syrup “ABC-
Dm” From two Identical well-filled flasks of the
syrup, samples were taken.every fifth day during a
period of 35 days One sample was placed m a refrl-
gerator at 3”, while the other was allowed to stay
on the laboratory bench, exposed to hght No special
Table 5 Effect of different solutions on the recovery of
ascorbic acid from a vitamin preparation (IDO-C)
Stored
Vitamin Sample before
sample solution analysis, hr mgltablet
I 8% acetic acid, oxygen-
free + 0 5% oxalic acid
II 8% acetic acid, oxygen-
free
III 8% acetic acid
3006
42 8
63 6
626
574 3t
,O”-
0 10 20 30
days
Fig 3 Stabdity of ascorbic acrd m a pharmaceutical syrup
(ABCDm) Upper curve stored m a refrigerator at 3°C
Lower curve stored on a laboratory bench at room
temperature
precautions to exclude ax were taken for any of the
samples The results are colle&d m Fig 3
In most of the existing methods for the assay of
ascorbic acid m pharmaceutical products, special
tlme-consummg procedures must be undertaken to
allow for each kmd of substance which can interfere
The results m Table 4 indicated that no components
of these samples would mterfere but m order to be
quite sure we made a critical study of nearly all the
constituents present m these preparations which
could possibly mterfere, and which have caused trou-
ble for many other mvestlgators
The experiments were divided into two parts. In
the first set the constituents were analysed alone and
m the second together with a known amount of ascor-
bic acid None of the constituents tested caused any
observable mterference. The lower hmlt for the
amount of added substance m each analysis was set
at about 1 mg The followmg substances were studied
raffinose, lactose, maltose, sucrose, fructose, galactose,
mannose, xylose, manrutol, mcotmamlde, calcium
panthothenate, retmol, thmmme, nboflavm, pyrldox-
me, cyanocobalamm, cahferol, tocopherol, fohc acid
and blotme There are also some other constituents
m the preparatlcins which we have not extensively
tested However, analyses on the whole preparations
but with ascorbic acid removed did not m&cate any
interferences
Several ways have been outlmed for the determma-
tlon of ascorbic acid m the presence of copper and
Iron salts Chapman and co-workersz4 have exammed
the Influence of copper and uon salts m eight different
frequently used methods and only one proved to be
quite satisfactory As some salts of copper and u-on
may also be expected to interfere with our method
It was necessary to find ways to prevent mterference
Ascorbic from these salts According to Hume and Kolthof?’
acid sodium citrate forms a stable complex with copper
found,
We performed some experiments m which a citrate
solution was added to the cell Just before the analysis
of a sample contammg ascorbic acid and cupric sul-
10 522 phate, and for small amounts of copper the method
worked well However, the use of EDTA instead of
10 522
citrate as a complexmg agent was found to work
10 490
excellently m all tests performed Ferrous Iron, which
 Determmatlon
occurs m many pharmaceutical preparations, does
not mterfere at all, which can be seen from the analy-
sts of the preparation “Multtplex Comp N”, which
contams about 15 mg of Fe’+ m each tablet Ferrous
n-on is, however, easily oxidized to ferric iron which
reacts with iodide and also, more seriously, with
ascorbic acid In order to prevent these reactions the
addition of fluoride or EDTA for complexation of
the ferric ions has proved satisfactory
CONCLUSION
A rapid and very accurate method has been devised
for the assay of ascorbic acid The method can be
used either m routme analyses or for the analysis of
completely unknown samples without any special
preparations, because the analysis current can be
reversed during a determmatton and an additional
amount of iodme generated if a sample contams more
ascorbic acid than expected Coloured or cloudy sam-
ples present no problems We have not found any
interfering substance, m the different samples ana-
lysed, of which the effect could not easily be eh-
mmated
Acknowledgement-The author wishes to thank Dr
Torsten Erlksson, Ferrosan AB, for kmd help with the
pharmaceutzal preparations and for valuable discussions
The author IS also indebted to Mrs Lena Andersson and
Miss Lena Lundman for valuable assistance with the ex-
periments
REFERENCES
1 J Tdhnans, P Hirsch and W Hirsch, Z Untersuch Z.e-
hewn, 1932, 63. I
of ascorbic acid 993
2 Oficral Methods of Analysis, 9th Ed, AOAC, Wash-
u$on, D C, 1960
3 Pharmacoaoela of the Urnted States. XVI. Mack Prmt-
mg Co, iaston,“Pa , 1960 ’
4 J H Roe and C A Kuether, J BzoZ Chem, 1943,
147, 339
5 M Schmall, C W Plfer and E G Wolhsh, Anal
Chem, 19.53,25. 1486
6 M Schmall, C W Plfer, E G Wolhsh, R Duschmsky
and H Gamer, rbld, 1954, 26. 1521
I C E Weeks and M J Deutsch, J Assoc O#ic Agr
Chem , 1965, 48, 1245
8 M H Hashml M A Shahid, M A Akhtar and N
A Chughtal, Mlcrochlm Acta, 1973, 901
9 M R F Ashworth, 7itrlmetrlc Orgamc Analyszs, Part
I, p 254 Interscience, New York, 1964
10 J W Stevens, Ind Eng Gem, Anal Ed, 1938, 10,
269
11 Pharmacopoeia of the Umted States, XVIII, Mack
Prmtmg Co, Easton, Pa, 1970
12 R Karlsson and K J Karrman, Talanta, 1971, 18, 459
13 L G Torstensson, lbzd, 1973, 20. 1319
14 R Karlsson and L G Torstensson. tbtd. 1974.21. 945
15 K S V Santhanam and V R Krishnai, Anal &em,
1961, 33. 1493
16 R Karlsson, Talanta, 1972, 19. 1639
17 L Erdey and E Bodor, Anal Chem, 1952 24. 418
18 G G Rao and V N Rao, Z Anal Chem, 1955, 147.
338
19 P V Krlshnamurty and K V Girl, J Indian Chem
Sot, 1941, 18, 1941
20 H Tauber and I S Klemer, J Blol Chem, 1935, 108.
563
21 R Strohecker and H M Henning, Mtamzn-Bestrm-
mungen, Merck, Darmstadt, 1963
22 M H Ha&ml, Assay of Vltamms m Phar-
maceutrcal Preparations, Wiley, New York, 1973
23 J D Pontmg, Ind Eng Chem , Anal Ed, 1943, 15.
389
24 D G Chapman, 0 Rochon and J A Campbell, Anal
Chem, 1951, 23, 1113
25 D N Hume and I M Kolthoff, Ind Eng Chem, Anal
Ed, 1944, 16, 103