682
INDUSTRIAL AND ENGINEERING CHEMISTRY
Satisfactory carbon dioxide for purging has also been prepared by slowly blowing off through a reducing valve a t least one half of the liquid contents of a full cylinder of carbon dioxide. However, the success of this procedure depends largely on the punty of the carbon dioxide as it is received. Preparation from solid carbon dioxide has been found to be more rapid and to give a uniform, low blank. The oxygen required for the analysis is preferably prepared in a totally enclosed electrolytic generator with a capacity of a t least 3 liters of oxygen a t atmospheric pressure and operated a t a pressure of 2 to 5 pounds per sq. inch. I n a totally enclosed generator an automatic valve is required in the hydrogen vent line in order to maintain equal pressures in the hydrogen and oxygen chambers and thus avoid surging of the electrolyte. If a totally enclosed generator is not used and the hydrogen is allowed to escape freely
Vol. 17, No. 11
to the atmosphere, there is danger that nitrogen from the atmosphere may dissolve in the electrolyte and diffuse into the oxygen chamber. All connections to the oxygen generator and. carbon dioxide cylinder must be gas-tight. Even though these lines are under pressure, nitrogen from the atmosphere often M u s e s in through very small leaks in valve packings, connections, etc. This difficulty has been a source of considerable annoyance. ACKNOWLEDGMENT
The authors are grateful to F. D. Tuemmler for his interest and encouragement. LITERATURE CITED (1) Baxter, G. P., and Hale, A. H., J . Am. Chem. SOC.,58, 510 (1936).
Determination of Sulfur Dioxide in Fruits J. D. PONTING AND GESTUR JOHNSON Western Regional Research Laboratory, U. S. Department of Agriculture, Albany,
The sulfur dioxide content of frozen fruits and of other types of fruit can be determhed rapidly b y extraction b y blending in buffered sodium chloride solution, which stabilizes sulfur dioxide against enzymic and autoxidation; filtration; treatment with alkali to dissociate combined sulfur dioxide; and acidification and titration with iodine, with and without added formaldehyde, which binds sulfur dioxide.
SULFUR
dioxide is commonly used to preserve color and flavor of dried fruits and certain frozen fruits. Its use requires control within rather wide limits in dried fruits but within relatively narrow limits in frozen fruits. As a result of the recent large expansion of the frozen-fruit industry, the problem of controlling sulfur dioxide concentration in frozen fruits has become more important. Methods of determination used satisfactorily for dried fruits have not proved adequate for frozen fruits. Sulfur dioxide in fruits is commonly measured by distillation from an acid solution into iodine or hydrogen peroxide solution, followed by measurement of the amount of oxidizing agent reduced. Distillation methods are slow, require rather elaborate equipment if many samples are t o be analyzed, and often give erratic results. For these reasons a rapid simplified method that yields accurate results for small amounts of sulfur dioxide is needed. Several methods not involving distillation have been developed. Iokhel’son and Nevstrueva (a) devised a method that is essentially as follows: The crushed sample is allowed to stand 30 minutes in an alcoholic potassium hydroxide solution, with frequent shaking .to extracq the sulfur dioxide; pigments and cellulose are precipitated with magneslum bisulfate; excess iodine is added to aliquots of the clear extract in the presence and absence of formaldehyde; and the residual iodine is titrated with sodium thiosulfate. Formaldehyde forms an aldehyde-bisulfite complex with sulfur dioxide that is stable to iodine titration, thus providing a blank for reducing substances other than sulfur dioxide. Bennett and Donovan (1) measured the sulfur dioxide content of citrus juices by liberating bound sulfur dioxide with strong sodium hydroxide, acidifying, and titrating directly with iodine. Sulfur dioxide is bound with acetone or removed by boiling, as a means of obtaining a blank value. Prater and co-workers (7) measured sulfur dioxide in dehydrated fruits by grinding in a food chopper blending in water, allowing the blended sample to stand 20 minutes in strong sodium hydroxide solution t o liberate bound sulfur dioxide, and titrating directly with iodine. They used acetone a t pH 2 to 3 t o obtain the blank value.
Calif.
A11 these methods leave much to be desired, especially when applied to frozen fruits. The extraction procedures (2,7 ) are not applicable t o fruits sampled when frozen. Clarification with magnesium bisulfate is tedious and back-titration with thiosulfate gives high and variable blank values on unsulfured fruit (2). In the authors’ experience the use of acetone to bind sulfur dioxide (1, 7 ) has given a poorer end point than formaldehyde. None of the methods is applicable to frozen fruits that contain active oxidizing enzymes and oxygen (as they usually do), because precautions are not taken to eliminate enzymic oxidation of sulfur dioxide during grinding snd extracting, and autoxidation during alkaline treatment. In the method presented here errors due to oxidation of sulfur dioxide have been eliminated or greatly reduced. The efficiency of extraction and clarification has been improved by the use of a blender and filtration, respectively. Formaldehyde is used to obtain a blank value for reducing substances other than sulfur dioxide as in the method of Iokhel’son and Nevstrueva ( 2 ) ; and bound sulfur dioxide is liberated with alkali as in the methods of Jensen (S), Bennett and Donovan ( I ) , and Prater et al. ( 7 ) , with alteration of conditions for alkaline treatment. The whole procedure requires about 15 to 20 minutes for one sample and much less if a series is analyzed. The method used for dried fruits varies somewhat from that used for sulfured fresh or irozen fruit, because of differences in enzyme and tissue oxygen content. METHODS
FORFRESH OR FROZEN FRUITS.A 100-gram sample of fruit is weighed into a blender (such as the Waring Blendor), 10 ml. of 0.5 M tartrate buffer a t pH 4.5 (tartaric acid and sodium hydroxide) and 490 ml. of aqueous sodium chloride solution, 20% by weight, are added, and the mixturejs blended 3 to 5 minutes. Sufficient blended material for subsequent iodine titrations (100 mm. or more) is filtered through coarse paper with suction or through a cotton milk filter disk. Filter aid can be used without loss of sulfur dioxide and is useful with gummy fruits such as apricots. A 50-ml. portion of filtrate is pipetted into each of two 125-ml. Erlenmeyer flasks, and to each flask 2 ml. of 1 N sodium hydroxide are added. After about 30 seconds the samples are acidified with approximately 2 ml. of 6 Ai hydrochloric acid. One flask is titrated a t once n i t h standard 0.02 N iodine, with 1 ml. of 1% starch solution as indicator. To the other flask approximately 1 ml. of 40% formaldehyde is added. This sample is allowed t o stand 10 minutes and is then titrated with iodine. The amount of sulfur dioxide in the fruit in parts per million on the fresh-weight basis is calculated from the difference in
ANALYTICAL EDITION
November, 1945
amount of iodine used in the presence and absence of formaldehyde as follows: For a 50-ml. aliquot,
Bisulfite Solutions during Blending for 10 Minutes
+
SO2 in p.p.m. = ml. of iodine (normality) (3800)
(2)
FORDRIEDFRUITS.A 20-gram sample of dried fruit (or 40 grams if sulfur dioxide is low) is blended in a mixture of 10 ml. of 0.5 M tartrate buffer at pH 4.5 and 490 ml. of water, for 10 minutes. Tough fruit should be cut into strips or ground before blending; soaking is not recommended, since it does not reduce the required blending time appreciably. After filtering, 50-ml. aliquots are titrated as for fresh or frozen fruits, except that 4 ml. of 1 N sodium hydroxide instead of 2 ml. are used t o liberate bound sulfur dioxide. The amount of sulfur dioxide on the “as is” basis is calculated from Equation 1; in this case the volume of liquid in the sample is calculated as the weight of sample X (% moisture/100). The moisture need be estimated only t o 5 t o 10%.
Table
I. Loss of Sulfur Dioxide from Aqueous Solutions during Blending
Solution Blended
1
.
NasSOa (PH 7.9) . NarSOs alcohol to 20% . . NazSOa 0.003% hydroquinone .. NaHSOr (pH 4.8) HaSOs (pH 1.0) lb:5
++
a
Minutes Blended 4 5 7 8 Loss of 902,% 6 7 . 1 9 1 . 0 . .. . . , .. 0.6 . 1 . 5 . , . 8.6 2
3
.
1.S 0.0
27.7a
..
.. ..
3.6 0.0
...
... 0.9 ...
1
... .., ...
0 23: 8
.. . . 1. ... 9 .. ..
11.3
..
Rapid corrosion of stainless steel blender blades.
DISCUSSION
OF METHOD
STABILIZATION OF SULFUR DIOXIDEDURING BLENDING. Preliminary experiments were conducted to ascertain which molecular species of the sulfurous acid group was most stable during blending. Aqueous solutions of sodium sulfite, sodium bisulfite, and sulfurous acid, each containing 770 p.p.m. of sulfur dioxide, were blended for varying lengths of time and the remaining sulfur dioxide was measured. The volume of solution blended in each case was 300 ml. With sodium sulfite solutions the losses were also measured in the presence of alcohol and hydroquinone as antioxidants. The results (Table I) indicate that the different molecular species have widely varying stabilities. It is evident that bisulfite ion is by far the most stable form, with loss during a 5-minute blending period of only about 1% without the addition of an antioxidant. Since the concentration of sulfur dioxide in the filtrate is usually not over 100 p.p.m. with the size of sample recommended, loss of sulfur dioxide is negligible during blending in buffer which maintains the sulfur dioxide mainly as bisulfite ion. Although bisulfite ion is the most stable of the molecular species, its stability varies markedly in different kinds of buffers at the same pH. I n order to find the most satisfactory buffer, sulfur dioxide losses were measured during blending for 10 minutes in various kinds of buffers and over a pH range of 4.0 to 5.5 (Table 11). In each test 380 ml. oi solution were blended. The data show that tartrate buffer was the most effective; consequently it was used in the extraction. The pH need not be maintained a t an exact value, since stability is practically constant between p H 4 and 5. A 0.01 111 concentration of buffer is used because the large volume provides an adequate amount at this concentration. Mitchell, Pitman, and Nichols (4) have shown that tartaric acid inhibits oxidation of sulfurous acid. A 3- t o 5-minute blending time is sufficient for soft fruits, but
II. Loss of Sulfur Dioxide from Variously Buffered Sodium
Table
Total SO2 in p.p.m. = ml. of iodine (normality) (640)volume of liquid in sample 500 (1) weight of sample The volume of liquid in fresh or frozen fruil may be considered as 94 ml. per 100 grams, and the formula becomes:
683
Buffer
PH 4.5 5.0 5.5 Loss of SOa, % 0.3 0.4 0.3 6.1 0.6 0 . 4 1.4 62.2 7.9 5.1 5.5 i:5 0.6 8.7 17.8 26.2 9 9 . 5 , 4.0
M tartrate M phthalate M oxalate M succinate M acetate No buffer (HCI NaOH)
0.1 0.1 0.1 0.1 0.1
+
...
3.3
1.5
0.9
1.1
PH 4 . 0 4 . 5 5 . 0 5 . 5 Loss of Soz, p.p.m. 0.6 2 2 14 3 2 9 335 . . . 36 33 38 9 3 4 8 . .. 104 152 350 . . . 23
10
6
6
for tough dried fruits a 10-minute period or even more should be used. It has been found with both the distillation and titration methods that there is no loss of sulfur dioxide from dried fruits, on blending within 15 minutes. The filtrate from the blended mixture is somewhat cloudy, but further clarification is not necessary; the cloudiness does not interfere with direct titration. PREVENTIOX O F ENZYMIC OXIDATION OF SULFURDIOXIDE. Blending of fruit in which active oxidizing enzymes are present causes a large loss of sulfur dioxide if buffer alone is used. This loss is evidently due to enzyme-catalyzed oxidation, a phenomenon apparently not heretofore recognized. The reaction is very rapid in both fresh sulfured fruit and in frozen fruit. When 300 p.p.m. of sulfur dioxide were added t o fresh apricots or other fruit having high phenol oxidase activity (the enzyme largely responsible for fruit browning), immediate blending in buffer and analysis revealed only 100 to 200 p.p.m. of residual sulfur dioxide; the remainder had been oxidized. When a known amount of sulfur dioxide was added to blanched fruit or fruit completely penetrated with sulfur dioxide, the added amount was recovered completely (Table 111). Since the texture of many fruits prevents penetration of sulfur dioxide much below the surface before or during frozen storage, a large portion of the enzyme below the surface remains potentially active. When the fruit is thawed the active enzyme is able to catalyze first the oxidation of sulfur dioxide and then (if all the sulfur dioxide is oxidized) oxidative darkening of the fruit. For example, fruits with higher phenol oxidase activity require a higher content of sulfur dioxide t o protect the color of the fruit when thawed. The texture of some fruits, such as apricots and peaches, a p pears to prevent entrance of oxygen, as well aj sulfur dioxide, into the fruit. This texture characteristic prevents enzyme activity in fruit having sulfur dioxide on the surface, since atmospheric oxygen is required for oxidation by phenol oxidase, and the surface enzyme exposed to oxygen has been inactivated.
Table
111.
Recovery by Titration Method of Sulfur Dioxide A d d e d to Fresh and Frozen Fruit Original Sot Total $01
Fruit
Apples SOz-treated SOz-treated Soptreated SOz-treated Steam-scalded Fresh. untreated
Extracting Solution Buffer Buffer Buffer Buffer Buffer Buffer Buffer Buffer 0.6 satd. NaCl (20%)
+
iSki”’
Cherries, untreated
1.5 satd. 5.9%) 1.7 satd. 3.6%) satd. (33.4%) satd. (33.4%)
+ +
Buffer NaCl Buffer NaCl
SO2 P.p.m.
Added
Found
P.p.m.
P.p.m.
0
24 70 45 24 146 498 265
50 94 73 199 143 497 240
0
265
263
0
94
92
0
94
95
0
94
94
0
94
86
0
160
159
24 30 30 173 0 0
6e4
INDUSTRIAL A N D ENGINEERING CHEMISTRY
However, when the fruit is frozen and thawed the texture is broken down, so that oxygen comes in contact with the active enzyme, allowing oxidation of sulfur dioxide and usually also darkening of the fruit. I n apples the texture is more open, allowing both oxygen and sulfur dioxide to penetrate; the enzyme is usually inactivated before the fruit is frozen and oxidation does not occur \Then it is defrosted.
PH Figure 1.
Effect of pH on Dissociation of Bound Sulfur Dioxide in Frozen Apples
If either fresh or thawed frozen fruit containing both sulfur dioxide and active oxidizing enzymes is allowed to stand at room temperature for a short time, sulfur dioxide will be lost by oxidation before analysis is started. Under present commercial packing conditions, frozen sulfured fruits containing active enzymes are frequently encountered; it is therefore necessary that frozen samples be analyzed for sulfur dioxide, by any method, before they begin to thaw. Furthermore, after sampling it is necePsary to inactivate the enzymes quickly. In the distillation methods this inactivation is accomplished with hot acid. In the proposed method, a fairly concentrated (20%) sodium chloride solution is used. On blending the sample in a salt solution of this concentration containing tartrate buffer in addition, no loss of sulfur dioxide occurs (Table 111). The optimum salt concentration lies between 17 and 23%. Below 17% the enzyme is not inactivated sufficiently rapidly and completely; above 23% there is some loss of sulfur dioxide, apparently due to frothing. USE O F FORMALDEHYDE FOR BLANKDETERMIXATIOS. With sulfur dioxide, formaldehyde forms an addition product, which is stable to oxidation by iodine. The reaction is rather slow, not being complete in less than about 8 minutes when a relatively large amount of sulfur dioxide is present. Increasing the concentration of formaldehyde does not appreciably increase the reaction rate. Iokhel’son and Nevstrueva (3) used 3 ml. of fonnaldehyde for an aliquot containing the sulfur dioxide from 2.5 grams of fruit; they do not mention a period of standing before titration. The authors have found 1 ml. of formaldehyde and a reaction period of 10 minutes sufficient to bind completely the sulfur dioxide in 2.5 grams of very highly sulfured fruits (containing 3500 p.p.m.). An aliquot containing 2.5 grams is usually taken in the present method for dried fruits. I n unsulfured fruits tested there has been no significant difference between titration values in the presence and absence of formaldehyde, thus indicating that formaldehyde does not bind any iodinetitratabIe substances other than sulfur dioxide. Experiments with pure ascorbic acid, probably the most concentrated natural reducing substance in fruits, showed that it was not bound at all by formaldehyde under the conditions used. Bennett and Donovan (1) used acetone to bind the sulfur di-
Vol. 17, No. 11
oxide and Pra,ter et al. (7) used acetone at pH 2 to 3. However, the addition product of sulfur dioxide and acetone does not appear to be as stable t o iodine as the formaldehyde addition product; the end point of the titration fades rather badly and is more difficult t o duplicate than when formaldehyde is used. An attempt was made to determine reducing substances, other than sulfur dioxide, with iodine in an acidified aliquot that had been boiled, as in the alternative method of Bennett and Donovan (1). However, with pure solutions it was not found possible to remove all the sulfur dioxide by boiling for less than about 20 minutes; this procedure was therefore abandoned. DISSOCIATION OF COMBINED SCLFURDIOXIDE. Jensen (3) and Bennett and Donovan (1) have shown that sulfur dioxide combines very rapidly with sugars, as with other aldehydes, the combined form being stable to iodine. Thus Bennett and Donovan found that when 1000 p.p.m. of sulfur dioxide were added to a sirup containing 39% invert sugar and 4% citric acid, more than 50% of the sulfur dioxide was in the combined form after 2 hours, iodine titration indicating only 467 p.p.m. However, the combined form is easily broken by the addition of alkali and will even dissociate gradually by itself if the equilibrium between free and combined sulfur dioxide is shifted by removal of the free compound with iodine or by distillation. Jensen determined sulfur dioxide in glucose sirups by adding 20 ml. of 5% sodium hydroxide to 50 grams of sirup and 50 grams of water and letting the mixture stand 15 minutes before acidifying and titrating with iodine. Bennett and Donovan allowed the alkali treatment to proceed 10 minutes in stoppered flasks. They stated that alkaline treatment for not more than 10 minutes has little effect on the natural reducing substances present. The authors have found that 30% or more of the ascorbic acid present may be lost during alkaline treatment, although this loss is mostly or entirely compensated for in the blank, provided it receives the same alkaline treatment. Two other factors make the alkaline treatment cited (1, 3) undesirable for fruits. First, treatment of 10 to 15 minutes’ duration has been found unnecessary, since dissociation of bound sulfur dioxide is complete in 30 seconds or less. Second, the use of strong alkali has been found to cause a significant loss of sulfur dioxide in frozen fruit filtrates. However, with 20% sodium chloride present, the loss is negligible. In the absence of salt the error due to oxidation of sulfur dioxide in a too strongly alkaline solution is fairly constant in each flask, regardless of the concentration of sulfur dioxide. For most frozen fruits the error is about 50 p.p.m. Since experience has shown that 75 to 100 p.p.m. of residual sulfur dioxide are sufficient to prevent oxidation in most fruits upon defrosting, an error of 50 p.p.m. in analysis cannot be tolerated. Dissociation of combined sulfur dioxide a t a pH below 11 does not cause appreciable oxidation of sulfur dioxide, in the short dissociation period employed. The effect of pH on dissociation of bound sulfur dioxide in a filtrate from frozen apples is shown in Figure 1. The curve was obtained with a dissociation period of 30 seconds. The dissociation is shown to be complete at pH 9 and nearly complete even at pH 7. Sodium hydroxide was used but sodium carbonate or trisodium phosphate is equally effective in this pH range. In strong alkali at a pH above 11 the sulfite ion is very easily oxidized. The loss of sulfur dioxide occurs in less than 30 seconds (in the absence of salt) but does not increase much with time, indicating that the sulfite ion is quickly oxidized by dissolved oxygen but that atmospheric oxygen does not diffuse rapidty into the solution. The fact that dissolved molecular oxygen is responsible for the oxidation is further shown by the following experiments: I n the f i s t experiment a filtrate of frozen sulfured apples waa analyzed with alkaline treatment a t pH 9.5; the sulfur dioxide content was 330 p.p.m. Then the same filtrate wm analyzed with alkaline treatment a t pH 11.7, before and after removal of most of the oxygen by evacuation and flushing with nitrogen,
November, 1945
ANALYTICAL EDITION
The undeaerated sample contained 281 p.p.m., while the deaerated sample contained 319 p.p.m. In the second experiment a filtrate of blanched frozen apples containing added sulfur dioxide was analyzed before and after oxygenation for 30 seconds. With alkaline treatment a t pH 11.7 the sulfur dioxide content was 424 p.p.m. before and 240 p.p.m. after oxygenation. With alkaline treatment at pH 10.7, the analysis %-as413 p.p.m. after oxygenation. In blanched fruit the oxygen content of the tissue is very low, causing little loss of sulfur dioxide even in strong alkali, when no oxygen is added. Evidently unblanched sulfured fruit (Experiment 1) contains a considerable amount of oxygen that remains in the filtrate. The action of salt in preventing oxidation of sulfur dioxide is twofold: Blending frozen fruit in a 20% salt solution prevents enzymic oxidation when active enzymes are present; it also prevents or greatly decreases autoxidation when a strong alkaline treatment is used to liberate bound sulfur dioxide. The data above show that at pH 10.7 the sulfur dioxide is stable in the absence of salt, even when the solution is oxygenated: apparently the critical pH value, above which autoxidation of sulfur dioxide is greatly accelerated, is above 11. For this reason it may be desirable to liberate bound sulfur dioxide from frozen fruit at a pH below 11, although this is not essential in the presence of salt. In order to avoid the necessity of making pH measurements, salt is used in the extracting solution for all fresh or frozen fruits, thus permitting the use of a constant amount of alkali. In dried fruit the oxygen content appears to be insignificantly low and the use of strong alkali does not cause an appreciable loss of sulfur dioxide even in the absence of salt. COMPARISON OF DIRECT TITRATION METHOD WITH OTHER METHODS
The two methods most commonly used for sulfur dioxide determination in food products are those of Nichols and Reed (6) and Monier-Williams (5), the latter being the official A.O.A.C. method. The Kichols and Reed method is extensively used on fruit products but as it is commonly used, no blank on unsulfured fruit is subtracted from the value obtained on the sulfured fruit; because unsulfured fruit of the same kind is usually unavailable. Severtheless, this method gives a blank value that varies for different fruits and is often high. Blank values of over 200 p.p.m. of sulfur dioxide have been obtained with unsulfured dried apricots, and 27 to 74 (average 53) p.p.m. with frozen blanched apples. When an average blank correction of 53 p.p.m. was made, a fairly good agreement between this method and the new direct-titration method was obtained when applied to frozen apples of varying sulfur dioxide contents. The direct-titration met.hod does not give a blank value on unsulfured fruit, either frozen or dried. The Monier-Williams volumetric method was found to give negligible blank values when untreated frozen fruits were distilled, but blank values of 60 to 200 p.p.m. of sulfur dioxide have been obtained n-ith unsulfured dried apricot's. Prater et al. ( 7 ) also obtained considerable blank values with various unsulfured dehydrated vegetables. Because of the insignificant blank value obtained by the Uonier-Williams volumetric method on untreated frozen fruit, this method \vas considered much more reliable as a check than the Sichols and Reed method. However, the blank value obtained by the Monier-Killiams volumetric method on untreated dried apricots and the lack of agreement betn-een the volumetric and gravimetric procedures aroused suspicion as to the accuracy of the volumetric method when applied t o dried fruit,. Therefore the gravimetric Monier-Williams method was used as a standard for dried fruits, since by this method a negligible blank value was obtained. In some cases the Monier-Williams method \\-asmodified for an added check on the titration method. This modification consisted in replacing the hydrogen peroxide solution in the receiver with 100 ml. of 0.1 M tartrate buffer (pH 4.5) and titrating the sulfur dioxide Kith 0.1 iV iodine solution. Tartrat,e buffer had
685
previously been found to stabilize sulfur dioxide, and recovery data on pure solutions containing 100 p.p.m. of sulfur dioxide indicated 99 to 100% recovery. However, when this procedure was applied to fruit, it was found necessary t o treat the buffered distillate with alkali to dissociate bisulfite presumably bound by aldehydes which distilled over from the fruit. The best procedure found was to acidify the buffer solution strongly with hydrochloric acid, titrate the free sulfur dioxide, make the solution alkaline to phenolphthalein, reacidify, and titrate the liberated sulfur dioxide. This procedure was repeated once more or until there was only approximately 0.05-ml. increase in titration. The main advantage of this procedure is that the products of distillation which are swept over with carbon dioxide are not in contact with a strong oxidizing agent for a long period of time as in the Monier-Williams methods. I n the volumetric SlonierKilliams method lower aldehydes may be oxidized to acids by the peroxide in the receiver and titrated as sulfur dioxide. Samples of fruit filtrates containing 20% sodium chloride could not be analyzed by this method, because hydrochloric acid distilled over and gave a blank value of up to 100 p.p.m. of sulfur dioxide. I t is not impossible with such a drastic treatment as hydrochloric acid distillation that in some cases volatile sulfhydryl compounds could be distilled over and oxidized to sulfate, especially when the acid solution is heated to boiling for barium sulfate precipitation. Experiment has shown that hydrogen peroxide is actually capable of oxidizing hydrogen sulfide to sulfate. Because of these possible errors in the Monier-Williams methods, it was thought advisable to compare the new method with the modified distillation method as well as with the volumetric and gravimetric variations of the Monier-Williams method. Some results of thcse comparative analyses are presented in Tables IV, V, and \-I. The
Table
IV. Comparative Determinations of Sulfur Dioxide in Frozen Fruits
Fruit
Direct titration
Cherries
214
Apricots -4pples
... ... 459 141 290
176 182
Method of .Inalyais Monier-Williams Monier-Williams Distillation volumetric= gravimetric into buffer Parts p e r million ... 220 224 317 322 ... ... 138 120 292 296 169,' i 9 6 160, 161 ... 454 460 ... 180 ... 187, 186 ... 210,' i 8 3
a Monier-Williams volumetrio method could not be used on samples blended with salt, because of high blank values on pure salt solution.
Table
V.
Fruit Apricots
Comparative Determinations of Sulfur Dioxide in Dried Fruits Direct titration
2iiO 2130 2360 2000 1900 1900 1780 1735 2060
Method of Analysis MonierMonierWilliitms Williams volumetric gravimetric Parts p e r million 2920 3200 2590 2770 2400 2430 2365 2510 224: 2280 2270 2260 2290 2190 2020 1740 1885 1980 2200 2240
1310
Distillation into buffer
.. 2240 2450 2220
..
..
686
INDUSTRIAL AND ENGINEERING CHEMISTRY
Table VI. Precision of Direct Titration and Monier-Williams Methods for Determining Sulfur Dioxide in Dried Fruits Method of Analysis Replicate Samples
Apples
No. 1 No. 2
No. 3 No. 4 Standard deviation (precision)
Direct titration
984
394 378 386
8
Monier-Williams Monier-Williama Volumetric gravimetric Parts per million
610 613 617
531 480 510
4
26
..
..
data were obtained for all methods on the same blended filtrate to ensure uniform sampling. The data for frozen fruit in Table IV show that very close agreement can be obtained among the different methods, the direct titration method agreeing even more closely with the MonierWilliams gravimetric method than does the Monier-Williams volumetric method. The comparative data for dried fruit in Tables V and VI, on the other hand, indicate the great difficulty in obtaining results of indisputable accuracy by any of the methods. With the Monier-Williams gravimetric method as a standard, the direct titration method usually gave lower results, while the MonierWilliams volumetric method often gave high results. However, when the blank value obtained on unsulfured fruit was subtracted, the latter also gave low results. I t would appear that the drastic treatment given the fruit in distilling it for 1.5 hours from hydrochloric acid solution dissociates a small fraction of sulfur dioxide which is too tightly bound to be dissociated by alkali
Vol. 17, No. 11
in the direct titration method. However, it is questionable whether this fraction is effective as a preservative when it is too tightly bound to be liberated by alkali. The fact that the direct titration method gave very consistent results on uniform samples of dried fruit (Table VI) and that the results were independent of time of alkaline treatment or pH (above a pH of about 10.5) indicates that there is a sharp dividing line between the fraction of sulfur dioxide dissociable by alkali and that dissociable only on prolonged boiling in acid. I n view of the high reproducibility with which the direct titration method measures the alkali-dissociable fraction, and the short time required for the determination, the authors believe the method should be useful even though it might yield somewhat lower results than the Monier-Williams method. The direct titration method seems to be more trustworthy than the Monier-Williams volumetric method, in view of the widely varying results yielded by the latter method on uniform samples. The recovery of sulfur dioxide added to frozen fruit, whether blanched, sulfured, or untreated, is shown in Table I11 to be accurate by the direct titration method. No recovery experiments were performed with dried fruits, as it was felt the results would lack significance, because of lack of time for combination of sulfur dioxide as in usual storage conditions. LITERATURE CITED
(1) Bennett, A. H., and Donovan, F. R., Analyst, 68,140 (1943). (2) Iokhel’son, D. B.,and Nevstrueva, A. I., Voprosy Pitaniya, 9, 25 (1940). (3) Jensen, H.R., Analyst, 53, 133 (1928). (4) Mitchell, J. S.,Pitman, G. A., and Nichols, P. F., IND.ENQ. CHEM.,ANAL.ED.,5,415 (1933). ( 5 ) Monier-Williams, G. W., “Determination of Sulfur Dioxide in Foods”, Reports on Public Health and Medica3 Subjects, No. 43,British iMinistry of Health, 1927. (6) Nichols, P. F., and Reed, H. M., IND.ENG.CHEM.,ANAL.ED.. 4. 79 (1932). (7) Prater, A. N.; Johnson, C. M., Pool, M.F., and Mackinney, G., Ibid., 16, 153 (1944).
Polarographic Determination of Nickel in Steel and Nickel O r e PHILIP W. WEST AND JAMES
F. DEAN*
Coates Chemical Laboratories, Louisiana State University, Baton Rouge, La.
A rapid, accurate method for determining nickel in steel and nickel ore has been developed. The method is based on the use of the polrrograph with sodium fluoride serving as the supporting electrolyte. When nickel i s present in quantities ranging between I and 570, accuracies of approximately 1 % can b e expected. The method compares favorably with spectrographic techniques.
T
H E analysis of steels and nickel ores for nickel content is ordinarily accomplished by application of the dimethylglyoxime reaction. Either gravimetric or colorimetric procedures may be used, depending on the accuracy required and the time available, Spectrographic methods are often used, particularly in the analysis of steels. The purpose of this paper is to present a method of analysis based on the use of the polarograph. The method is rapid and the accuracy compares favorably with the more lengthy gravimetric procedures. The cost of equipment is much less than that required for spectrographic analyses and the accuracy is greater. 1 Present address, Esao Laboratories, Standard Oil Company of New Jersey (Louisiana Division), Baton Rouge, La.
There are approximately thirty ions which are reduced a t the dropping mercury electrode when sodium fluoride is used as a supporting electrolyte. A later paper will give more detailed discussions of the general behavior of various ions in this supporting medium. The present paper is restricted to information dealing specifically with the determination of nickel in materials of high iron content such as steel and nickel ore. Sodium fluoride is especially useful in this case because it serves to remove large amounts of iron in the form of a crystalline precipitate which separates from acid as well as alkaline solutions. The danger of coprecipitation in acid medium is slight. Small amounts of iron are sequestered by sodium fluoride as the fluoride complex, which is so stable that the reduction potential of the iron is shifted to such a negative value that it does not give a wave within the limits of the discharge potential of the sodium or hydrogen ions present in the supporting electrolyte. This is contrary to the behavior described by Kolthoff and Lingane (1, page 277) who give a value of -1.36 volts vs. the saturated calomel electrode for the half-wave potential of ferric iron in 0.04 to 0.8 hl potassium fluoride, this value being attributed to Stackelberg and Freyhold.