Microdetermination of Methylenedioxyl or Combined Formaldehyde Groups MORTON BEROZA Entomology Research Branch, Agricultural Research Service,
U. S.
Department of Agriculture, Beltsville,
Md.
used. Bricker and Johnson (2) give a method of purifying the technical product. Sulfuric acid, c.P., concentrated and 10M. Trioxane, C.P., or formaldehyde. Aqueous solution of known concentration containing approximately 30 y per ml. Thirty micrograms of trioxane are equivalent to 30 y or 1 micromole of formaldehyde. Trioxane should be weighed in a closed container, as it is slightly volatile. Its solutions are stable for a t least 2 weeks. Apparatus. Test tubes, 22 x 175 mm., made of borosilicate glass with a 24/40 female ground joint and glass stoppers to fit. The 50-ml. level is permanently marked (etched, ceramic ink, or by any other means). Adaptor consisting of a 24/40 male ground joint connected to a stopcock (Corning Catalog No. 9120), which is connected by means of rubber tubing to a water-pump aspirator and a manometer. All measurements were made on a Beckman Model DU spectiophotometer using cuvettes 1 cm. square. A constant slit width of 0.03 mm. was employed. Modified Procedure. The sample to be analyzed should contain 1 micromole of the compound, unless it is known that more than one methylenedioxyl group is present, in which case a proportionately smaller amount of compound should be taken. Make up a weighed sample to a definite volume with redistilled acetone or another easily volatile solvent. Introduce an aliquot of this solution (usually 0 . 5 or 1 ml.) into the bottom of the test tube and attach i t to the adaptor with the stopcock open. Evaporate the solvent under reduced pressure while swirling the bottom of the tube in a beaker of water at room temperature. The manometer will indicate when all the solvent is evaporated (about 1 minute with acetone). Close the stopcock and detach the test tube. Add 1 ml. of the chromotropic acid solution and follow with 5 ml. of concentrated sulfuric acid from an OstwaldFolin pipet with continuous shaking Swirl vigorously to be
A good method for the determination of methylenedioxyl or combined formaldehyde groups is needed for studies of chemical structure. The method of Bricker and Johnson for the determination of combined formaldehyde with chromotropic-sulfuric acid has been modified to obtain results that more closely approach theoretical values. By the modified procedure an absorbance equivalent to 0.75 to 1.08 moles of formaldehyde is obtained per mole of methylenedioxyl group. The method has been tested by its application to fi4 selected compounds of known structure. It is rapid and accurate, and the results are reproducible.
T
HE methylenedioxyl group is necessary for the synergistic
action of sesamin with pyrethrins (9). This group is found in alkaloids, cancer-inhibiting substances ( I O ) , and other physiologically active compounds of natural origin. A good determination for this grouping is needed for studies of chemical structure. Methods for the determination of the methylenedioxyl group are based on its hydrolytic cleavage by mineral acids to form formaldehyde, which ia determined either gravimetrically or colorimetrically in the same acid medium. Phloroglucinol ( 6 , 8 ) , carbazole (18), phenol (16), and resorcinol (13) have been used as formaldehyde precipitants; Schiff’s reagent (11) and chromotropic acid (2, 14) have been used to determine formaldehyde colorimetrically. The gravimetric methods are more timeconsuming and require much more material than the colorimetric procedures. The use of chromotropic acid in the presence of strong sulfuric acid to detect minute amounts of formaldehyde was originated by Eegriwe (6). This highly specific test was developed into a quantitative method almost simultaneously by 11acFayden (14) and Bricker and Johnson ( 2 ) . The latter investigators also suggested the use of their method for combined formaldehydethat is, for methylenedioxyl groups, They listed six compounds that gave practically theoretical yields of formaldehyde (1 mole per mole of methylenedioxyl group), the implication being that quantitative yields could be obtained from all methylenedioxyl groups. In a later paper Bricker and Vail ( 3 ) stated that the formaldehyde of constitution, as in safrole and piperonal, might be determined by the procedure of Bricker and Johnson ( 2 ) . However, by this method only 0.50 mole of formaldehyde was obtained per mole of safrole. Other compounds Rere also found to give low results. This method was therefore in need of improvement. The method described is a modification of Bricker and Johnson’s. It gives results that approach the theoretical more closely than do those obtained by the unmodified method. Interferences in the method by labile methylene groups (production of formaldehyde in the presence of acid) have also been studied. It is known that methylene groups attached to nitrogen interfere (18); methylene groups attached to sulfur may also interfere, as will primary alcohol groups under certain conditions. The present method should be applicable for the determination of labile methylene groups and of compounds containing these groupings. The method is rapid and the results are reproducible.
Table I. LMolesof Formaldehyde Liberated from Compounds Containing Methylenedioxyl Groups Number Unof CHzOn Modified modified Groups Procedure Procedure
Compound 3-Buten-1-01, 1-(3,4-rnethylenedioxyphenyl)m-Dioxane, 4-(p-methoxyphenyl)5-methylm-Dioxane, 5-methyl-4-(3,4-methylenedioxyphenyl) m-Dioxane, 4-phenyl Isosafrole &Mannitol. 1.3:2.5:4.6-trimethvl-
-
Piperine Piperonal Piperonylic acid Protopine Safrole d,Z-Xylitol, 2 , 4 :3,5-dimethyleneAverage yield per compound, % Average deviation from true value,
1.00
1.01
0.87
1.00
0.91
0.85
2.00 1.00 1.00
2.02 0.93 0.79
1.91 0.75 0.51
3.00 1.00 1.00
3.24 0.96 1.07
...
1.00 1.00 1.00 1.00 1.00 2.00 1.00 2.00
1.02 1.05 0.87 1.03 1.05 2.08 0.75 2.15 98
1.01
70
8
...
...
...
0.76 1.03
1.05 ...
0.50
... 82
16
Table 11. Moles of Formaldehyde Liberated from Compounds Containing Methylenedioxyl Plus Interfering Groups
EXPERIMENTAL
Chromotropic acid (lJ8-dihydroxynaphthalene3,6-disulfonic acid), Eastman Kodak P-230. The solid or a solution containing LOO mg. per ml. (prepared fresh daily) is Reagents.
1970
Compound Asarinin Piperonyl alcohol Piperonyl alcohol, benzoate Podophyllotoxin Sesamin Sesarnol
Number of CH?02 Groups 2.00 1.00
Interfering Group Ether CHnOH
1.00 1.00 2.00 1.00
CHIOH, ester Lactone Ether Phenol
UnModified modified Procedure Procedure 2.40 ... 1.57 1.21 1.25 1.33 2.50 0.57
0.93
...
1.73
0 47
V O L U M E 26, N O . 12, D E C E M B E R 1 9 5 4
1971
__..___
GO-y level.
Table 111. Moles of Formaldehyde Liberated from Compounds Not Containing Methylenedioxyl Groups Compound Interfering compounds: Acrylaniide, N,N’-methylene-
Interfering Group
bisN-CH-N hnisyl alcohol CHzOH Anisyl alcohol. benzoate CHIOH, ester Benzyl alcohol, p-chloroCHIOH Benzyl alcohol. p-chloro-, benzoate CHzOH, ester Cinnarnvl alcohol CHzOH Glucose CHzOH S=CHz Glycinonitrile, S-methylene Hexamethylenetetramine N-CHz-Ti Pinoresinol. acetate Ether a-Trithiane S-CHI-S Vanillyl alcohol CHIOH Vanillyl alcohol, dibenzoate CHzOH, ester Veratryl alcohol CHZOH o-Veratryl alcohol CHzOH Soninterfering compounds: rlcetic acid, 2,4-dichlorophenox~ Anisaldehyde Amsir nrid Benzyl-%ohol Benzyl alcohol, o-chloroDihydrorotenone 2,O-Formoxylidide Furfural Furfuryl alcohol (;unnidine, nitrate 2.4-Hexadien-1-01 !-Histidine, dihydrochloride Prntaerythritol 3-Pm e.. n.-~ l. - o. l. . r~ . ~
I’lienol. 2,4-dichloro1.9-Propanedio1, 2-amino-2(hydroxvinethv1)1.3-Propanediol, -2-(hydrouyinethyl)-Z-nitro1-Propanol, 3-phenylRotenone Sorbitol Sparteine, sulfate Urea Uric arid Vanillin Vanillylamine, hvdrochloride o-Veratraldehyde Veratric acid
UnModified modified Procedure Procedure
...
0.77 0.85 0.71 0.25
0.57 0.51 0.18
0.19 0.15 0.25 0.87 1.10 0.79
0.17 0.15 0.20 0.92 1.35
2.66
,..
...
0.78 0.04 0.75 0.68
0.55 0.62 0.47 0 67
0.00 0.01 0.00 0.01 0.06 0.07 0.01 0.05
0.01 0.00 0.00 0.00 0.06 0.12 0.00 0.02 0.055
0.065
0.01 0.03B 0.00 0.12n
... ...
... ...
0.00 0.00 0.00 0.00
0.00
0.00
0.01 0.01 0.09 0.01 0.035 0.00 0.01 0.01 0.03 0.04 0.00 “ Commercial sample. If purified through t h e tetraacetate, it foiriialdehyde. Communication by C . E . Bricker.
0.01 0.01 0.15
...
...
0.00 0.01 0.01 0.01 0.04 0.00 gives no
sure the acid has reached all the sample, and place the test tube uncovered in a boiling-water bath for 30 minutes. Immerse in cold water for a few minutes and make up to the mark (50 ml.) with 10M sulfuric acid. Mix the contents of the tube by inserting the glass stopper and inverting four or five times. After the solution has cooled to room temperature, determine its absorbance a t 570 mp against a reagent blank similarly treated. At the same time determine the absorbance developed by 1 ml. of the solution containing the known amount of trioxane or formaldehyde, but do not evaporate this solution, and add 100 f 1 mg. of solid chromotropic acid in place of the solution. Should the absorbance of the unknown he more than that produced by 37 y of trioxane or formaldehyde (Beer’s law is followed), take a Fmaller quantity of the compound so that the absorbance is in the range of that produced by 30 zk 7 y. If the compound is insoluble in acetone or other volatile solvent, or if it is itself volatile in the above-described evaporation, dissolve it in water, take an aliquot, and make i t up to 1 ml. with water. Add 100 mg. of the solid chromotropic acid and then treat as outlined above, starting with the addition of the concentrated sulfuric acid. To save material or because of solubility difficulties, weigh out nonvolatile compounds on a microbalance in a platinum boat and introduce the boat directly into the test tube for analysis. Solubility of compounds rarely causes difficulties, since only 0.1 to 0.5 mg. is generally used per analysis. Unmodified Procedure. Follow the procedure described above but add only 50 mg. of chromotropic acid instead of 100 mg. per determination and use a sample twice as large. Results. The results obtained by both procedures on compounds containing methylenedioxyl groups are given in Tables I and I1 and those obtained on compounds not containing such groups in Table 111. DISCUSSION
I n the analysis of formaldehyde or trioxane by the unmodified procedure good agreement with Beer’s law was obtained to the
I t was also found that increasing the chromotropic acid concentration did not give much increase in color. These findings are in agreement with the previous work ( 2 ) . On the other hand, in the analysis of combined formaldehyde liberated by the acid hydrolysis, increasing the chromctropic acid concentration frequently gave a significant increase in color, and the smaller the sample the more closely the results agreed with the theoretical. The results obtained with isosafrole are typical of those giving low results by the unmodified procedure. The effect of sample size of isosafrole in determinations containing 100 ing. of chromotropic acid is shown below Sample,
&I$.
0.012 0.024 0.048 0.060 0,084
Mole Formaldehyde per Mole Isosafrole 0 81 0 79 0 72 0 67 0.64
The smaller the sample the closer the amount G f formaldehyde liberated approached the theoretical value of 1 mole for isosafrole, although this decrease was accompanied by a lesser precision. One micromole of methylenedioxyl group (theory is 30 y of formaldehyde) gave adequate precision along with results that were in good agreement with the theoretical. The effect of varying the chromotropic acid concentration in the determinations on isosafrole is shoivn below. Chromotropic Acid per Detn., M g . 50 75 100
125
Absorbance 0.368 0.402 0,446 0.469
The absorbance developed from a constant sample size (0 060 mg. ) increased as the weight of chromotropic acid per determination was increased. Good results were obtained with a minimum of 100 mg. I n doubtful analyses larger amounts of chromotropic acid may be tried to improve the results. In these cases the reults should he based on a reagent blank and standard formaldehyde solution which contain the same amount of chromotropic acid. Although the changes made in the modified procedure appear to be minor, Table I shows that they give a marked improvement in results. The recovery of formaldehyde per compound averaged 08% by the modified procedure as compared u ith only 82% by the unmodified one. The compounds that liberated practically theoretical quantities of formaldehyde by the unmodified procedure liberated the same or only slightly larger quantities by the modified procedure (see piperonal and piperonylic acid). Compounds that gave much less than theoretical yields by the unmodified procedure gave yields that were much closer to the theoretical by the modified procedure (see safrole and piperine). With the modified procedure the maximum absorbance per mole of methylenedioxyl group was obtained from the methylene derivatives of Iylitol and mannitol, which gave an absorbance equivalent to 1.08 moles of formaldehyde; the minimum absorbance was obtained fiom safrole, which yielded 0 75 mole. The absorbance of compounds that gave practically theoretical yields of formaldehj de (1.00 to 1.08 moles) showed only minor deviations from Beer’s law when determined a t several different concentrations. However, the conformance of a compound with Beer’s law should not be taken for granted because it gives a theoretical yield of formaldehyde, but should be checked. Compounds giving lower than theoretical yields of formaldehyde usually gave results that were affected by concentration. Quantitative analyses on such compounds require calibration curves. The color developed by most of the compounds was checked about 18 hours after color formation. S o instance of appreciable change in absorbance was found. I n several experiments compounds were introduced in methanol and ethyl alcohol solutions without subsequent evaporation of the
.
ANALYTICAL CHEMISTRY
1972 solvent (no water added) and were analyzed. The absorbance developed in each case was only slightly different from that obtained with an aqueous solution. Bricker and Johnson have indicsated that these solvents do not interfere in the determination of formaldehyde itself. INTERFERENCES
To establish the validity of the modified procedure, compounds
of known structure were tesbed and interferences were found. Compounds similar to those exhibiting interferences were analyzed in order to determine the origin of these interferences arid to indicate a means of ovcrcoming them. Because compountls were selected on the basis that they might interfere, Table I11 contains a disproportionately larger number of compounds showing such interfence than is ordinarily the case. Labile Methylene Groups. Methylene groups attached to nitrogen are reported always to liberat'e formaldehyde (18). Thus X-methyleneglycinonitrile, lV,AV'-methylenebisacrylamide, a n d hexamethylenetetramine do so. The formaldehyde produced by the last compound is markedly affected by sample size, larger yields per mole being generat.ed by smaller samples. N o forni:ildehyde is produced by --S=CH--S= (histidine dihydro(urea), =S-CHzCH*--N= chloride, uric acid), =?;-CO--S= (sparteine sulfate), =N--CSH--?;= (guanidine n h a t e ) , or formyl (2,6-formosylidide) groups. YIrthylene groups at'tached to /S-CHp, sulfui. may liberate formaldrh\~d(~, s-Tri t,hinrie,H2C s,
's--c
H/
gives 2.66 moles. Primary Carbinol Groups. The following discussion refers to t h e results obtained by the modified procedure. Piperonal, piperonylic acid, and piperonyl alcohol contain one methylenedioxyl group each. Although the first two compounds yield approximately 1 mole of formaldehyde, piperonyl alcohol gives 1.57 moles. The excess formaldehyde originates from the primnrj- carbinol (-CH20H) group, since it was possible to obtain from 0.68 to 0.85 mole of formaldehyde per mole of vanillyl, misyl, veratryl, and o-veratryl alcohols even though they do not contain any methylenedioxyl groups. If one considers fornialdehyde to be methylene glycol-that is, its monohydrate (IIO--CH20H)-the reaction map be represented as follows:
Apparently the bond bt>twwn t!w carl~inoland the aromatic ring is a iyeak one. Aromatic or aromatic-type carboxylic acids are quantitatively decarboxylatrd simply by heating thcin in quinoline in the presence of a catalyst ( 1 ). The present situation, in which a carbinol instead of a carboxyl group is split off, is analogous. However, under t,he present conditions bmzyl alcohol does not yield formaldehyde, whereas the substituted benzyl alcohols do. The substitutions on the aromatic portion of the ring seem to weaken the bond between the carbinol and ring sufficient.ly PO that the carbinol group is hydrolyzed as illustrated abovr. The yields of formaldehyde from 0- and p-chlorobenzyl alcohols (O.OG and 0.25 mole) show how positional isomerism may affect the result,s. Cinnamyl alcohol yields 0.15 mole of formaldehyde, whereas 3-phenyl-1-propanol yields practically none. The double bond appears to contribute to t'he weakening of t,he bond holding the carbinol group. However, 2,1-hexadien-l-ol gives only 0.035 mole. The double bond in the 3 position, as in 3-pentenol, gives negative results. The present color reaction with different concentrations of reagents was used by Klein and Weissman ( l a ) to determine glucose in blood serum. The method permits the quantitative estimation of hexoses and hexosedisaccharides in the presence of pentoces. The authors showed that formaldehyde results from
5-hydrcx~methylfurfural,which is produced from hexoses but not from pentoses in the presence of strong acid. Interference should therefore be expected from all sugars that produce this compound. This is another example of the primary carbinol group responding to the present reaction. Furfuryl (an aromatic-type) alcohol, like benzyl alcohol, gives practically no formaldehyde, but the, addition of the aldehyde group to the furfuryl alcohol, as in 5-hydroxyniethylfurfural, aids in the hydrolysis of the primary carbinol group. Glucose produces 0.25 mole of formaldehyde. Sorbitol, a related compound, produces practically none. Pentaerythritol, 2(hydroxymethyl)-2-nitro-l,3-propanediol, and 2arnino-2(hydrox~methy1)-1,3-propanediol give practically none. The determination of methylenedioxyl groups in sesamin ant1 asarinin gave 2.50 and 2.40 moles, respectively .(theory, 2.0 moles). The reason for the high results on these optical isomer:: was clarified when it was found that pinoresinol acetate, a simi1:tr compound containing the same central nucleue but no methylenediovyl groups, gave 0.79 mole. The formulas of sesamin, usarinin, and pinoresinol acrtate follow:
CH, \ '0'
CH--~R, /
RI and It2
R1
=
= O Z C H(methylcnedioxyl) ~ for sesamin and asarinin OCOCH3, RP = OCH, for pinoresinol acetate
The tetrahydrofuran rings are d i t in the presence of strong acid and probably form primary carbinol groups which upon hydrolytic scission produce formaldehyde. The production of 1.33 molrs of formaldehyde by podophyllotoxin was explained on the basis of one methylenedioxyl and one lactonc group. The opening of the lactone group by the ncitl exposes a primary carbinol group. If the primary carbinol group is oxidized to an aldehyde or acid, the product no longer liberates formaldehyde and the int,ei,ference is eliminated. For examole, the vanillyl, anisyl, and veratryl alcohols liberate formaldehyde, whereas v:tnillin, anisaldehyde, anisic acid, o-veratraldehyde, and veratric acid do not. If the carbinol is esterified, less formaldehyde than that liberated by the original compound is obtained (vanillyl dibenzoate, anisyl benzoate, p-chlorohenzyl benaoak, and piperon!-I benzoate). Other Interferences. The weed killer 2,4dichlorophenox.~-acetic acid gives a color similar to t h a t produced by formaldehyde when heated to 150" C. for 2 minutes with a chromotropicsulfuric acid solution ( 7 ) . Wit'h the present procedure no color is produced by this acid or by 2,4-dichlorophenol. Vanillylamine does not interfere, Speck ( 1 6 ) used the present reaction to determine diacetyl quantitatively. Rotenone and dihydrorotenone, two coniplrs highly oxygenated compounds, do not interfere. Phenol (16), resorcinol ( I S ) , and phloroglucinol (6)react with formaldehyde to produce a precipitate. Phenols compete n-it.h chromotropic acid (also a phenol) for the formaldehyde liberated. Thus sesamol gave unsatisfact.ory results 10.57 mole, also offcolor) because of the existence of a competing reactiori-namely, that of a phenol (sesamol) and formaldehyde. This type of rcaction (sesamol and furfural in the presence of strong sulfuric* acid) has been used to make quantitative estimates of sesamol (4,1 7 ) . The use of 250 mg. of chromotropic acid per determination gives somewhat better resulh (0.70 mole). Phenols should therefore be expected t o interfere with the present determination, although increased amounts of chromotropic acid may partially overcome this interference. Bricker and Johnson have mentioned color inhibitors and
V O L U M E 26, NO. 12, D E C E M B E R 1 9 5 4 other interferences. These must be present in relatively large amounts and will not interfere with the present determination, in which only the compound (usually 0.1 to 0.5 mg.) is present. However, any attempt to analyze for compounds containing methylrnedioxyl groups in t.he presence of foreign material must take into account this source of interference. ACKNOWLEDGMENT
The author is grateful to W.F. Barthel, S. I. Gcrtler, M. S. Schechter, and Martin Jacobson, of the Entomology Research Branch, and Lyndon F. Small and J. L. Hartwell, of the National Institutes of Health, for supplying compounds used in this study. LITERATURE CITED
(1) Beroza, XI., AIVAL.CHEM.,25, 177 (1953). (2) Bricker, C. E., and Johnson, H. R., IND.ENG. CHEM..AIV
