A Spectrophotometric Study of the Schiff Reaction as Applied to the Quantitative Determination of Sulfur Dioxide ROBERT V. NAUMAN, PHILIP W. WEST, AND FRANqOIS TRON Department o f Chemistry, Louisiana State University, Baton Rouge, l a .
G. C. GAEKE, Jr. Kem-Tech laboratories, Inc., Bafon Rouge, l a .
b A study was undertaken to elucidate the mechanism of the reaction between sulfur dioxide, pararosaniline, and formaldehyde. Parallel studies were made of reactions involving simple amines such as aniline. By comparison with known ultraviolet and visible spectra, various reactions can be postulated and the final reaction product can b e established to be a sulfonic acid (R-NHCHZ-SO~H). The studies are significant because of the wide use of this reaction in the determination of sulfur dioxide in air pollution studies.
F
tlic Schiff reaction has been uqed for cahnracterizing aldehydes by their color reaction with magenta decolorized by sulfur dioxide. The same type of reaction may be utilized :IS a test for suliur dioxide and/or sulfite. The test n a i fi Stcigmann (6) n h o u t wid-bleached fuchsin solution and formnldchyde as a reagent for the qualitative identification of iulhtes. Grant is), Atkin ( I ) , Crone and Hoggs (9). and K e s t and Gaekc (11) h a ~ eused rimous modifications of thi- reaction for the detr.rmination of sillfur dioxide in tlic atmoq)heie. The nicchanirm of the Schiff reaction has been studied frequently, and a t least three mechaiiisms appear in the literntnre, these are: Formation of a Srhiff b a v ( 7 ) o K M A N S E-EARS
It-SI-I:
+
HCHO
-
+
+
-
H-XH-SOzH
+ HCHO
R--SH--SO?-CH&H Forinntion of a sulfonic. acid ( 5 ) It--"?
t
SO1
T
EXPERIMENTAL
Apparatus. I3cckman bfodcl DK recording spectrophotometer. Sargent hIodel T' Oscillometer. Reagents. Pararosaniline hydrochloride,., Eastman Kodak Co. Formaldehyde. 407, in water, analytical reagent grade. Aniline. All other chemicals and reagents were prepared from :inalytical reagent grade chemiral?. The solutionh for tlie spectrophotometric studies were prepared as follow : One millilit'cr of 0.04y0 pararosaniline hyclrorhloride solution was added to 1 ml. of 0.2% aldehyde, followed by the addition of a few drops of a freshly prepared solution of sodium sulfite or of sulfur dioxide st,andardixed by iodometry. The solutions n-ere made up to 10 nil. with distilled water. These approximate concentrations were present) in tlie solutions: pararosaniline, l O - ~ M ; formaldehyde, 10-3M; HCI. 10-2M; sulfur dioxide, 10-5 to 10-4M.
SO?
R--S=C€I? H,O Formation of an amino wlfinic acid tollowed liy tlic addition of thc aldehyde (12) It-XHn SOs It--SH---SO?H
Tlie.;e and other possible mechanims h a w been considered in the present study. Cltravioltlt ant1 ~ l s i b l e absorption spectrophotometry have h e n uqed to study the reartion ~inder renditions similar to tho-? utilizetl by T e s t and Gaeke (11) in their qiiantitative procedure for the determination of wlfur dioxide. Under the-c conditioni the spectroscopic results bupport the mclcha n i m that inrliitlr+ the iyntheqis of R sulfonic' acid.
H,O HCHO ------+
Il--SH-CHz--SO,H
RESULTS AND D13CUSSION
Tlic clectronic absorption spectrum of apl)roximat,ely lO-"If aqueous pararosaniline in the TOO to 225 nip region is given in Figure 1. The spcctrum is characterized b>-t)hrcr%main absorption bands located at 538. 280. and 232 inp. The band a t 538 nip is typicsd of the lowest energ>- haiid of all triphenyl methane dyes and must be asqociated u-ith a transit,ion iuvolvinp the molecule as a whole in the form of a positive ion whic.h allons thc clrctrons in a11 three
lmizene rings and the unpaired c ~ h troni on the nitrogen atoms to interact through the central carbon atom. A typical structure contributing to one of the d a t e s involved i j
PI
Y
SH*
The band at 280 mp must bc associated with an elertronic transition localized primarily within the benzene rings and the associated nitrogen atom because of its close similarity to the highest energy band of aniline, the spectrum of which is shown in Figure 2. The band at 232 mp cannot be, nor need be in this case! definitely assigned, but t,he fact that it is less intense in pararosaniline than the 280-mp band, whereas in aniline it is more intense than the 280 band, leads one to believe that. it is associated primarily with a transition that involves electronic interaction through the wnt8ral carbon atom. Any reaction that prevents electronic interact.ion through the central carbon atom should, if this interpretation is correct, eliminate the 538-mp absorption band and thus decolorize tlic solution. This could be accomplished either by covalent bonding of' a fourth group to the central carbon atom or by lionding all three pairs of unsharcd elcl.ctrons on the nitrogen atoms. The occurrence of both of thcsr possibilities has been demonstrated. When sodium sulfite or sulfur dioxide a t B pH greater t'han 6 is added to the pararosaniline solution the 538-mp band disappears, the 280-mp hand decreases in intensity, and the 232-mp band increases in int'ensity and is red shifted slightly (Figure 3). 'This compound can be kept under a nitrogen atmosphere but it is unstable and senFitiw to oxidation when left exposed VOL. 32, NO. 10, SEPTEMBER 1960
1307
to air where it reverts to pararosaniline quickly, probably because of oxidation of the sulfite. It is assumed that bisulfite ion now present in rather high concentrat,ion has bonded to the central carbon atom t o give ”2
H
J
-
~
-
?
~
-
N
H
~
SOSH
The 28o-mp band decreases in iiiteiisity probably because of the decreased importance of the
--a=”I __
*
form when the net positive charge has been removed from the molecule. The relative intensities of tlic 280- and the 232-mp bands are similar to those in aniline itself, although the 232-mp band is red shifted to peak a t 245-mp. In pararosaniline. the configuration
W a v e Lenpth
Figure 1.
Action
(mpl
of HCI on pararosaniline
- 0.004%
---I.
dye in water 0.004% dye in 0.01 N HCI 0.004% dye in 0.05N HCI
01
NHz
is probably a minor contributing form to the states involved in the ,538-mp transition; but the XH?-
232
280 Wave
form is probably a major contributing form to the states responsible for the 280-mp transition. HoiTever, when bisulfite is added and the central positive charge is removed, repulcion of negative charges would make a conhguration n ith only one structurc ; a = X H z
+
f o r (~.\ample. XH* I
more probable. The other tn-o benzene rings would give toluene-like absorption in the 255-mp region and thus by 1308 *
ANALYTICAL CHEMISTRY
Figure 2.
superposition tend to shift the 232-mp band to the red as compared wit,h aniline. Sodium sulfite has no influeiicae on the aniline spectrum, just as one would have expected. When HCl is added to p:mrosaniline solutions, the solution is decolorized and all three bands disappear, to be replaced by a series of sharper bands centered around 255 nip. This is shown in part in Figure 1, but :i higher dye (:oncentration must be uwd to slioiv the new bands. When aniline is acidified, both the 280- and 232-mp bands disappear and again a new series of sharper bands centered around 255 mp appear (Figure 2). These hands, which appear a t approximately 255 mp, are w r y similar to bands of toluene ( 4 ) . All
length
(mp)
Action of HCI on aniline
three -et5 of bands (pararosanilinc~, :miline, and toluene) are essentially tho same and belong to the tolueno and toluene-like aiiilinium ions:
:inti ”2’
1
A ! j
\ /’
In tlic. hitter ion there is no interac-tion brt,\-vccvi b(wzrnc rings, and this mol-
Wave
Figure 3.
length
(mr)
Action of NaZS03 on pararosaniline
___ ---__-
Dye only Dye plus 5 pg. SO2 Dye plus 10 pg. SO?
ecule gives essentially the same spectrum as does toluene, diphenyl methane,
and triphenyl methane. A weak acid gives the same results as HC1 but, as expected, higher concentrations are required. The action of sulfur dioxide at a p H of less than 6 is that of a R-eak acid on both pararosaniline and aniline, being cnoniplicated only by a peak at 280 mp a t high sulfur dioxide concentrations, d i i c h is found also in pure sulfur dioxide solutioiis. Mixtures of sulfur dioxide and hydrochloric acid show only the acid effects. Formaldehyde b y itself has little effect on either the pararosaniline or aniline y)cctra. In dilute fornialdehydc
no discernible cliangc occurs in either case. In 1M formaldehyde pararosaniline peaks occur at 548, 285, and 235 mp, compared with 538, 280, and 230 mp in water. Thc aniline peaks in 1M formaldehyde occur a t 283 and 238 mp> compared with 280 and 230 in water. The shift of the peak positions is probably just the solvent shift (in acetone A, = 548 mp) due to changing dielectric constant, but might be due to formation of a weak compound such as a Schiff base, RX;=CH*, or a polymer, RNH-CH2-NHR. For both substances the bandy are somewhat >harper in 1M formaldehyde. Hydrochloric acid and sodium sulfite
1.0
0.0 Wave length
Figure 4.
-----
(mgl
Final reaction product Pararosaniline ( 5 pg. SO2) Pararosaniline (40pg. SO*) Aniline ( 1 00 p g . Sol)
affect the dye in the same manner in t,he presence of concentrated formaldehyde as they do in pure aqueous solution; however, slightly more hydrochloric acid is required to eliminate the normal dye absorption. This is shown more readily by the addition of formaldehyde to acidified dye that gives the toluenelike spectrum; with dilute formaldehyde no modification of the spectrum occurs, but a t higher formaldehyde concentrations the dye spectrum and the color reappear progressively with the same shift observed in formaldehyde solutions. There is only a slight change in p H (from 1.4 to 1.5) as the color reappears, but the color reappearance must be associated with the removal of some hydrogen ions b y the formaldehyde from the bleached dye to pi ve normal pararosaniline, When formaldehyde is added to a lmrarosaniline solution decolorized b y sulfur dioxide in neutral or basic s o h tion (HS03-), the pink color and the normal pararosaniline spectrum reappear immediately and then slowly partially disappear again to a degree determined by the amount of sulfur dioxide present. As in the absence of formaldehyde. the stability of the system depends on the absence of oxygen. The renppearance of the normal pararosaniline spect>rumis probably a kinetic result caused b y the rapid formation of a bisulfite addition compound that in part destroys the bisulfite bonded to the central carbon atom of pararosaniline. Equilibrium is then probably slowly established and some of the pararosaniline bisulfite is re-formed n i t h the resulting partial disappearance of the color anti the normal par:irowdinci spectrum. Formaldehyde. sulfur dioxitlt.. and pararosaniline in acidic solution. as used for the determination of sulfur dioxide, give a new irreversible product. A new compound, the spectrum of which is shown in Figure 4, results. The intensity bf this spectrum depends on the amount of d f u r dioxide present and is the basis for the quantitat,ive determination of sulfur dioxide. This compound i> purple in color but its spectrum is cscentially the siinic as that of pararosaniline except for :i retl shift; t)he bands occur at 560. 290, and 240 mp, coml~arcdn-ith 5%. 280, anti 232 nip. for Iwxrosaniline. ;Iniline, formaldehyde. a n d sulfur dioxidc in modpratrly acidic solution givc, a spectrum similar to that of aniline, but it is retl shifted; the bands of t.he new compound are :it 287 and 242 mp compared n-ibh t h e of aniline at 280 and 232 mp. This is essenti:illj- thv same differenre found between pararosaniline and the new purple compound. The compound ~ - S H - C H ~ - - S O ~ N ~ ~ VOL. 32, NO. 10, SEPTEMBER 1960
1309
by the same reaction that gave
was synthesized by classical procedures (2, 6). It was purified by recrystallization and its spectrum as determined in neutral and acid solutions. The spectrum was found to be identical n i t h that of curve 3, Figure 4, the spectrum of the aniline, sulfur dioxide, and formaldehyde mixture. It is then tentatively concluded that the reaction product is
Q--SH-CH,--SOJI
in moderately acidic solution. dqueoua 10-4hl aniline requires 10-zM HC1 for the elimination of the aniline hpectruni and the full appearance of thc aniliniuni ion spectrum, but 10-4.1f
in n a t e r requires 10-'M HCl for the elimination of its characteristic spectrum (Figure 4) and the full appearance of the toluene-like spectrum of
Thus, a tenfold greatw hydrogen ion concentration is required for changing the spectrum of the wbstituted aniline than for aniline alonc. The purple coinpound formcd from pararosaniline has appreciably different properties from those of pararosaniline. It requires 10iV HC1 to rliminate the normal tripheiiyl methane dye spectrum of a n approximately 10-5AMsolution of the purple compound compared with lO-'hl HCl to eliminate the normal 10-6M pararosaniline spectra; (.onsequently, excess sulfur dioxide will not eliminate the purple as it will the pink color of pararosaniline. The purple compound does not undergo the reaction of primary amines, that of diazotization and coupling with @-naphthol; pararosaniline undergoes this reaction easily. Both the purple compound and pararosaniline are easily decolorized and give spectra like that of aniline or that. of the pararosaniline bisulfi c shown in Figure 3 when sodium hydroxide is added to give a p H of 8 or higher. This indicates that the central carbon atom is equally easily accessible in both cornpourids to give uncolored products. The diffcrcnce must lie in the amine groups. An attempt was made to synthesizt.
which has a spectrum identical with the product resulting from the aniline, sulfur dioxide, formaldehyde, and acid reaction mixture. The synthesis gave the following: a stable insoluble green product that is soluble in HCl giving a purple solution n-ibh a triphenyl methane dye type spectrum with a first maximum whirh changes from 540 to 580 mp n-ith changing p H and which is almost decolorized by excess €IC1 to give a solution to which addition of sulfur dioxide givw the Schiff color reaction (probably a condensation product that, is hydrolized by tht, HCl to give back pararosaniline); and a very soluble violet compound that has a spectrum with maxima at 560, 292. and 240 mp nearly idcntic:il with that of the purple compound that results from tho pararosaniline. formaldehydr, riilfur dioxide, and HCl mixt>ure. This violet compound requires vcry conccntrated HC1 (approximately 1 O M ) to decolorize it and to obtain the toluene-like spectrum. Sodium hydroxide dccolorizes the compound to give the :miline-like spectrum easily. This violet compound should be the nnc sought in the syntliesil;.
13 10
0
ANALYTICAL CHEMISTRY
HCHO
+ HC1-
-r)-XH,
L
Cl-CH,OH
+ CICHZ-OH
-*
/CH,Cl -F)--SH,
+ HIO
CONCLUSIONS
Sulfur dioxide added to :I pararosaniline solution decolorizcd Ly HC1 in the presence of fornialdehydc causes the formation of a new purple compound, the 560-mp absorption band of which may be used for the quantitative determination of sulfur dioxide. The nature of the purple compound is in question. The purple rompound cannot be a Schiff base because no role 11-ould be played by sulfur dioxide if this were the case. The colorless compound that exist.. before the addition of sulfur
4
i
dioxide is not a Schiff base because spectral evidence indicates that the nitrogen atom takes part in four single bonds that cause the colorless compound to have a toluene-like spectrum. The purple compound cannot be an amino sulfinic acid because there is no spectroscopic evidence for the formation of such a compound. Such compounds are easily formed by the reaction of sulfur dioxide gas with dry amines, but tjhey arc not stable in aqueous solution. Chloronicthylation of pnrarosaniline catalyzed by hulfur dioxide was considered as a possible mrchanism for the production of the purple compound. Direct chloromethylation ( I O ) of pararosaniline by formaldehyde in concentrated HC1 gave a brown precipitate iiisolubk in sodium hydroxide. HC1, or organic solvents and a dark blue solution that had the 5SO-nip absorption band but' also had a n absorption band continuous between 300 and 210 nip. Bobh the prwipitatc and the blue solution failed to have the properties of the purple compound, hut chloromethylation
can be ruled out on other bases as well. The intensity of the 560-mp band of the purple compound depends on sulfur dioxide concentration ; the sulfur dioxide reacts as a component of the reaction rather than as a catalyst. The purple compound does not give the reaction of the primary amine which chloromethylation product should give. By thi. elimination of other choice5 and considering the positive spectral evidence, it ib concluded that the purple compound is a sulfonic acid derivative of pararosaniline :
SH-CHz-SO I I CH?
,H
c1-
This conclusion agrees with that of Rumpf (6). The mechanism of the reaction seems to be that the pararosaniline is decolorized by hydrogen ion to give a colorless aniliniuni salt
be assumed that the dichlorosulfitomercurate(I1) complex behaves exactly as sulfite ion in the rcsction. When, a- in the quantitative procedure for d f u r dioxidr, the 902 is collectrcl :i-the dichlorosulfitomercurate(I1) complex, [HgCISOa on13 the first btrp of the reaction, and thus only tht. reaction'. speed, i5 changed. Under thrse conditions the first rraction step is prohabl) [HgCl?SOj]- 2
+ HC'HO
11
+
~
(JH
HgCL
+6 S'OJH
and the rrmaintlrr of thc reaction> ale as indicated above. I n addition, otlirr aldehydes nil1 give iimilar results hut the final product i- less stable. 0
2
I
Figure 5.
HpC14'
Composition of HgC14-2-S02 complex
I
Cl
The formaldehyde has no effect on this reaction but combines with the sulfur (lioxide as proposed b y Rumpf (5):
HCHO
+ SO1
-
y'
HCH 'SO,
Evidence for the formation of a conipound of this type is given b y the observation that sulfur dioxide in acidic solution has a band at 280 nip which disappears when formaldehyde is added even at a p H as low as 1. This compound then combines with tlie amine (pararosaniline) salt t o give a sulfonic acid
R-XH-CHZ-SO~H
+ H+
or HOCH2-SO3H
SUMMARY
4
SO;/
+ R'-&
R--NH-CH,-SO,H
-+
+ H?O
which is the purple compound. The new colored product is coonsistent with the spectra observed in sulfonic acid compounds synthesized b y direct procedures. The color appears because the existing acid concentration is not
great enough to give the colorlesf anilinium-like salt of the sulfonic acid. Independent experiments show that both aniline and pararosaniline sulfonic acids require appreciably higher hydrogen ion concentrations for the formation of the colorless anilinium-like salts than do the unsubstituted aniline and pararosaniline. The composition of the mercury(I1)sulfur dioxide complex wab also iiivestigated. West and Gaelic in their original HCHO or
+ H?SOY
1
HO-CH?-SO,H publication on the use of t'etrachloromercurate(I1) as a n absorber for sulfur dioxide (11) indicated that thf complex formed was c~isulfitomerc~uratc (11) [ Hg(SO& J -*. Results of further investigation? on the coinpodition of this complex using the high-frequency Oscillometer to determine the molar ratio between mercury(I1) ion and sulfite ion are s1ion.n in Figure 5 . These data indicate that there is a 1 to 1 molar ratio b e t w w i mercury(I1) ion and sulfite ion. The htructure of tlie complex is t'herefore dichlorosulfitomercurate(I1) [HgC1?SOB]-* and not disulfitomercurate(I1) [Hg(S03)2]-2. The broad absorption band of mercury(I1) ion in thv ultraviolet region made it impossible to carry out the studies using the dichlorosulfitomercurate(I1) ion as u 'ource of sulfur dioxidc. Honever, since there is no shift in the position of the major band in the visible region and no difference in the intensity of this hand, it, may
Ultraviolet and visible absorption spectra indicate that when HC1 is added to pararosaniline, the decolorization occurs becauseof the formation of a trianiliniuiii c~hloromcthanewhich has an absorpt,ion qiectrum very similar to that of eitlipr toluene or triphenyl methane. both of which are colorless. The addition of sulfur dioxide in any available f o i m in the presence of formaldehyde lcatls to the synthesis of a pararosaiiiliiic niethylsulfonic acid which gives an ahsorption spectrum similar to that of painrosanilinc, but which requires a 1iight.r avid concwtration than d o ~ s uiisubstituted pararosaniline, for the form:ition of the colorless trianiliniuni form. As a rrsult of the above factors, the formation of the color of the 1i:irarosaniline methylsulfanic acid is considered to b t x the basis for the quantitative estimation of sulfur dioxide. LITERATURE CITED
(1) Atkin, S.,hsar.. CHEM.22,917 (1'330). ( 2 ) Backer, H. J., Mulder, I€.,Rec. trcra. chim. 5 2 , 451 (1933); 53, 120 (1934). (3) Grant; \y. AI., -%S.iL. CHE?f. 19, 345 (1947). (4) Jones. R.S . , J . Am. Chent. ~ O C .67, 2127 (1915): I)oub, Leonud, Vnndenbelt, ,J. AI.? [hid., 69, 2714 (1947). ( 5 ) Rumpf, P.,. l m . chinz. 3, 327 ( l W 5 ) . (6) Rumpf, P., Bull. soc. chim. 5 , 871 (1938). (7) Shoesniith, J. B., Sossorl, C. E., Hetherington, -1. C., J . Cheni. Soc. (Londoa) 1927,2221. (8) Steigmmn, .I.,J . Soc. ('hen. I d . 61,18 (1948). (9) Urone, 1'. F.,Boggs, \V. E., .\SAL. CHEX 23, 1517 (1951). ( l o ) \Tagnt,r, E. C., J . .4m. C h e w SOC. 55,78-1(1933). (11) \\'est. P. K.. Gaeke. G. C.. .ISAL. (!HEM. 28, 1816 (1956). ' (12) Wieland, H., dchearing, C.. Her. 54, \ -
2527 (1921). RECEIVEI)for rwicw- rlpril 1, 1960. =Iccepted May 1 1, 1960. Southxvest Regional IIeeting. ICY, Baton Rouge, La.. Deceriit)w 1959.
VOL. 32,
NO. 10, SEPTEMBER 1960
131 1
