VI11 from nor HN2.HB, VII, aould be comparable to that for the forniation o f IV from VI, the ion VI11 oncc formed n ould be e w n t i J l y compl(tc1ly changed to the ethyleneiminc IX a t liiglier pH. Thus the rate of t1cconii)o~itionof nor “2, a\ + l i m n in Figure G, increases from 1” 3 to -5 in a manncr comparable t o that for H N 2 but then decreaqcs to essentially zero a t about p H 8. From this decrease, it is apparent, as might be expected, that the ethyleneimine IX is rclatively resistant to nucleophilic attack. This significant characteristic difference in hydrolytic behavior of these two mustards may provide the basis for a simple analytical method by which secondary and tertiary amino nitrogen mustards may be measured separately in a mixture. R e propose to apply such a niethod to a study of the cyclization
and of the hydrolytic behavior of the series of new secondary nitrogcii niustarde reported earlier (6, 6 , 9)in rclation t o thvir biulogical propcrtics. ACKNOWLEDGMENT
The authors gratefully acknonledge their indebtedness to Schimon Schichor n h o carried out cssen i d l y a11 of tlic experimental work in this study. LITERATURE CITED
(1) Arnold, H., Bourseaux, F., Brock, N , Nature 181, 931 (1958). ( 2 ) Ausman, R. K., University of Win-
nesota Medical School, LIinneapolis, hlinn., private communication. (3) Bartlett, P. D., Ross, S. D., Swain, C. G., J . Am. Chem. SOC.69, 2971 (1947). (4) Colowick, C. P., Kaplan, N. O., “Methods in Enzymology,” 1701. 1, pp. 140, 142, Academic Press, New Tork, 1955.
( 5 ) Epstein, J., Rosenthal, R. W., Ess, R.J., .4KAL. CHEM.27, 1435 (1955). (6) Friedman, 0. &I., Boger, E., J . . 4 ~ , Chem. SOC.78, 4659 ( 1956). (7) Friedman, 0 . >I., Seligman, A A r , , Zbid.,76, 655 (1954). (8)Zbid.,p. 658. (9) Friedman, 0. hf., Somnier, 1?I Boyer, ., E., Zbicl.,8 2 , 5202 (19GO). (10) Golumbic. C.. Fruton. J. S., Berg518, 536
( l i ) Ogsten, A. G., Holiday, E. R., Philpot, J. St L , Stocken, L. A,, T r a n r . Faraday SOC.44, 4,5 (1948). (12) Rutenburg, A. hl., Persky, L , FriPdman, 0. hl., Seligman, A. XI.,J . Pha, m. Esptl. Therap. 3,483 (1954). RECEIVED for review October 26, 1960 Accepted March 9, 1961. Work supported in part by a research grant from the Sational Institutes of Health, U. S. Public Health Service, S o . CY-2130 and by a contract with the Cancer Chemotherapy National Service Center, National Cancer Institute, U. S. Public Hrnlth Service, No. SA-43-ph-3075.
Separate Spectrophotometric Determinations of Hydroquinone, Benzoquinone, and the Monomethylether of Hydroquinone in Acrylic Monomers DELWIN P. JOHNSON and FRANK E. CRITCHFIELD Development Deparfment, Technical Cenfer, Union Carbide Chemicals W. Va.
b Colorimetric methods have been developed for separately determining (HQ), benzoquinone hydroquinone (BQ), and the monomethylether of hydroquinone (MMHQ) in acrylic monomers. HQ i s oxidized to BQ, followed b y reaction of the latter with 2,4-dinitrophenylhydrazine. The resulting hydrazone i s then reacted with a mixture of diethanolamine and pyridine to produce a blue color which i s proportional to the B Q concentration. Free BQ i s determined b y omitting the oxidation step. Concentt ations ranging from 2 to 40 p.p.m. of either HQ or B Q can b e determined with an accuracy of 1270, and highly reproducible results can b e obtained for as much as 100 p.p.m. with reasonable acciracy. M M H Q is determined a t concentrations ranging from 5 to 1000 p.p.m. b y a selective nitrosation technique. The intensity of the yellow color thus formed i s a function of the M M H Q concentration.
A
evaluation of the activity of acrylic monomers requires a complete knowledge of the amount and nature of additives and impurities present. Additives include THOROUGH
910
ANALYTICAL CHEMISTRY
Co., Division o f
inhibitors nhich are added in lon- concentratioiis to prevent polymerization. Among substances employed for this purpose are hydroquinone (HQ) and the monoinethylether of hydroquinone (MMHQ). Usually only one or the other of these inhibitors is used but in Some c a w both may be present. I n addition, the degradation products of the inhibitors may be present also. Chief among these is benzoquinone (BQ), the oxidation product of H Q Nunierous methods have been proposed for analyzing for these substances but most are severely limited by either sensitivity or specificity. Afanas’ev ( 1 ) described a colorimetric method for determining HQ but it is subject to interference from other phenolic compounds. Ghosh and Bhattacharya ( 3 ) used a nitration technique for determining HQ in the presence of BQ but this approach is also limited in specificity. Stephen, (5, 6) in two separate papers, described an indirect method for determining HQ in methacrylic acid. More recently Lacoste and associates (4) have described a highly sensitive method for HQ but, as written, it will not distinguish between H Q and BQ. I n general, attempts to determine NMHQ have been based on its phenolic
Union Carbide C o p , South Charlesfon,
properties. Such methods include active hydrogen determination and nitrosation techniques. However, HQ usually undergoes these same reactions and, because it occurs as an impurity in commercial AlMHQ, has been a major source of interference. The need for separate and accurate determinations of both H Q and AIAIHQ, as ne11 as BQ, prompted the development of the present methods. Free BQ is determined by its reaction n i t h 2,4-dinitrophenylhydrazineto producthe bis(dinitropheny1hpdrazone). Th, hydrazone is then reacted nith a mild base to produce an intense blue color. HQ is determined by first oxidizing it to BQ, follom-ed by a determination of the latter. These methods are applicable in the presence of any concentration of LlAIHQ. The monomethylether of hydroquinone (NlIHQ) is determined in the acrylates and acrylonitrile by a selective nitrosation technique which produces an intense yellow color. The method is applicable to these niononiers in the presence of both BQ and HQ. It has also been adapted to acrylic acid but in the absence of HQ. These methods can be used for both trace and assay analyses.
EXPERIMENTAL
Reagents. 2,4-Dinitroplienylhy(liazine. Dissolve 0.1 gram of reagent Rrade mntcrial in 50 nil. of carbonylflee nictlianol ( 2 ) containing 4 ml. of ronccritrated hydrochloric acid and dilute to 100 nil. nitli nater. Dietlianolaniine in _uvridine, 1% " (v. v.) solution. Procedure for HO and BO. 'I'ransfci 0.2 t o 1 nil. Gf t h e acrylate or aciyloniti~lc,11c ~ g h e dto t h e nearest 0 1 mg., i n t o cnch of two 25-ml. glassstop pe 1 ecl F:I ad 113 t ed cy lin d ers nu mbeled 1 silt1 2 A d d 2 nil. of n a t e r to each of these and also t o a third c3linder t o be procesed as a blank. To t h e blank and t h e fiist sample, a d d 1 ml. of 15% (v./v.) sulfuric acid. T o samplt nunihrr 2, add 0 5 ml. of 0.1N sodiiim carbonate and immcdiatcly snirl vigorous11 for 5 seconds. Quichly add 1 ml of 15% sulfuric acid and mix. Add 1 ml. of 2,klinitrophenylh) drazine solution t o cnch cylinder, stoppcr, and digcst in a a a t e r bath a t i o " i 2" C. for 1 hour. Shake the samplcs intermittently during this period. Cool thc samplcs to room temperature and add 13 nil. of aater. Add, Rith a pipet, sufficient benzene to make it total nonaqueous layer of 6 ml. Shake vigorously and allou the phases to separate. Pipet 2 nil. of the top layer into n test tube containing 10 ml. of the dietlianolnmine-pyridine solution. MI\ ncll and determine the absorbance of the wmples in 1-em. cells a t 620 mw using the blank to zero the spectrophotometer Dctcrmine the BQ concentration i n each from a pieviouqly 111iyni cd c,~lihutioncurT c. Calculations. in-~ sample 1 sample in no 1
pg of BQ . ~-
g
p p m BQ pg.
of BQ in sample 2 sample in no. 2
-~
g.
111
-ample
- p.p.m. BQ
=
p,p.m. HQ in sample
Procedure for M M H Q in Acrylates and Acrylonitrile. Transfer 10 ml. of t h e sample, weighed to t>henearest 0.1 mg., into a 50-ml. volumetric flask and dilute t o t h e mark with chloroform. Transfer 5 ml. of t h e solution into a 125-iiil. separatory funnel. T o another funnel, transfer 5 ml. of chloroform for a blank determination. Pldd 1 ml. of 0.1N aqueous sodium hydroxide t o each funnel, shake well, and immediately add 1 ml. of 3 5 hytlrochloric acid. Add 1 nil. of 2% sodium iiitrit,e and immediately shake for :~pprosimntely20 seconds. Draw tht. loner layer from the funnel into a t r s t tube and add approsimately 1 gr:tni of sodium chloride crystals to rmiove droplcts of n-ater. Decant the solution from the test tubes into 1-em. cclls and, using the blank t o zero the instrument, determine the absorbance of the sample a t 420 mp. Determine the M M H Q concentration by applying the nbsorbnncc to a calibration curve. S o t c : If the intensity of the color is bcyxid the rnngc of the instrument,
dilute t,lie sample nntl 11l;~ikwitli chloroform to a suitable volume and multiply the. results by the dilution factor. Procedure for M M H Q in Acrylic Acid. Dilute 10 ml. of t h e sample, weighed t o the nearcst 0.1 mg., t o csactly 50 nil. with diniethylformamide (DJIF). Pipet 5 nil. into a 25-1111.glass-stoppered graduated cylinder and add, ivith n pipet, 0.5 nil. of 3 N hydrochloric acid :\n(l 0.5 nil. of 2y0 sodium nitrite. 3Iix n.cl1 and allow to stand for 10 minutes a t rooni temperature. I'rocc~s 3 nil. of pure D h I F in the s a n e mariner to serve as a, blank. Using the blank to zero the instrument, detcrminc thc absorbance of the sample in 1-cm. rclls a t 420 nip. Detcrmine t h e Ml.raIHQ conceritration by applying the absorbance to n cnlibration curve. K o t c : If the iiitcnsity of the color is beyond the range of the instrument, dilute the sample a i d blank with D M F t o a suitable volume and multiply the results by the dilution factor. Calibration Curves. T h e calihration curve used in this investigation for HQ and BQ was prepared by analyzing 1-ml. aliquots of a series of butyl acetate solutions containing 5 to 40 pg. of HQ per ml. Butyl acetate was used because HQ-free acrylates are not readily available. The same curve was used for both BQ and HQ determinations. For MAZHQ in acry1;ttc:s and acrylonitrile, the calibration curve was prepared from standards consisting of hIMHQ dissolved in a 20% misture of butyl acetate in chloroform. Fivemilliliter portions containing 10 to 200 pg. of h/IXHQ were processed. The curve for MLIHQ in acrylic acid was prepared by processing acetic acid solutions containing 20 to 300 pg. of X d H Q per ml. .icetic acid was used because uninhibited glacial acrylic acid usually is not readily available. DISCUSSION
H Q and BQ. T h e reaction of 2,4dinitrophenylhydrazine with B Q is analogous t o the reaction of this reagent with simple aliphatic ketones. T h e product, however, is a bis(hydrazone) which is a n orange, benzenesoluble, substance. Upon treatment n.ith a base, the product fornis a n intense blue color. T h e mechanism of the color fornintion may be illustrated as follows:
"'r
-
~
--1
1
I
i 20
IO
30 ' SECONDS
40
' 2 50 '
60
Figure 1 . Effect of sodibm carbonate reaction time on recovery of hydroquinone The continuous conjugation through nine double bonds and the resonating character of the quinoidal ions probably account for the blue color. HQ is air-oxidized to BQ simply by treating the sample with sodium carbonate follolj ed by acidification with sulfuric acid. The sodium carbonate treatment m u 4 be 'carefully controlled to obtain reliable results. Figure 1 shows that if the treatment is extended beyond about 15 seconds there is B significant decrease in sensitivity. Therefore, the solution must be acidified R ithin this time if mauimuni sensitivity and accuracy are to be maintained. The diethanolaniinc-pyridine solution n-as selected as the color developing medium to avoid interference from other simple carbonJ-1 conipounds such as acetone. Afore basic media produce colors with some of these derivatives. Although these colors are usually red to wine red, they would, nevertheless, cause slight interference at the 620-mp wave length used to measure the blue BQ color. Therefore, by using diethanolamine, interference from these compounds is limited to the consumption of the 2,4-dinitrophenylhydrazine reagent. The blue color of the BQ derivative follom Beer's law for quantities up to approximately 40 pg. of UQ or HQ. Although the curve loses sensitivity beyond 40 pg., it is highly reproducible and as much as 100 p g S j of either compound can be determined with reasonable accuracy. The color, which actually appears green against the yellow background of the excess 2,4dinitrophenylhydrazine, is stable for at least 24 hours. Although the molecular weights of BQ and HQ differ by t n o units, the
-0
0 --
\\ N
J\
0
//
N 0-
J\
0
0-
VOL. 33, NO. 7,JUNE 1961
911
colors produced by equimolecular quantities of these substances are essentially identical in intensity. Therefore, the calibration curves may be used ivterchangeably. The sensitivity and precision of the method are demonstrated in Tables I and 11. The samples were prepared by mixing known quantities of BQ and HQ in the acrylates. One portion of the sample was analyzed directly for free BQ while another portion was subjected to the oxidation treatment, followed by a total BQ determination. The difference between these values represents the HQ concentration. The data in Tables I and I1 were obtained with I-cm. absorption cells. By using 2-cm. cells, the sensitivity can be increased to show an absorbance of approximately 0.02 unit per pg, of either BQ or HQ. Quinhydrone. T h e response of quinhydrone to the 2,4-dinitrophenylhydrazine method was checked. This substance is a n association product consisting of equimolecular quantities of B Q and HQ. T h e quinhydrone was found to react as unassociated molecules of B Q and H Q rather than as a single associated compound. M M H Q . T h e probable course of t h e nitrosation of M M H Q is illustrated in the following equation:
I.
Table
Determination of Known Concentrations of Hydroquinone in Acrylates
Acrylate Ethyl Ethyl Ethyl ZEthylhexyl 2-Ethylhexy l 2-Ethylhexyl
Table
II.
Added 6.7
20.6 33.6 8.1
24.3 40.6
Determination
912
e
ANALYTICAL CHEMISTRY
HQ, P.P.M. Found 6.6 21.1 33.1 8.1 24.8 40.0
of Known Concentrations of Benzoquinone in Acrylates
BQ, P.P.M. Added Found 7.8 7.8 21.7 22.2 28.7 29.2 43.4 42.7 11.3 11.9 22.5 22.6 29.9 30.7 40.6 40.0
Acrylate Ethyl Ethyl Ethyl Ethyl 2-Ethylhexyl 2-Ethylhexyl 2-Ethylhexyl 2-E thylhexyl
ratio and hence a high degree of accuracy. If the ratio is changed, a new calibration curve must be prepared a t the corresponding ratio. I1
1
CHa I
1
0
The yellow color produced thus can be attributed to the resonating character of the nitroso derivative. The color is stable indefinitely and conforms to Beer’s law over a wide range of concentrations. Maximum absorbance of the color occurs a t 405 mp; however, as shown in Figure 2, the blank produces a slight interference a t this wave length. The interference is due to the excess nitrous acid and is overcome by using a wave length of 420 nip. Although the latter point occurs on the shoulder of the curve, the absorbance is easily reproduced and the loss of sensitivity is negligible. The nature of the reaction medium has a direct influence on the intensity of the color. I n pure chloroform, for example, the intensity of the color is approximately 10% greater than that produced in the pure acrylate. I n DMF, the color is only about 65% as intense as that produced in chloroform. It is important, therefore, that the sample size be carefully controlled to maintain a constant sample to solvent
Recovery Deviation from Per cent average yo 98.4 -1.6 102.6 f2.6 98.6 -1.4 100.0 0.0 101.9 fl.9 98.4 -1.6 Av. 100.0 fl.5 _.
if present. BQ, however, does not react. Interference from low concentrations of H Q in the acrylates or acrylonitrile is, therefore, circumvented by oxidizing that compound to BQ. This is accomplished by the sodium hydroxide-hydrochloric acid treatment prior to the nitrosation. When H Q is present in concentrations greater than approximately 0.050/0, the oxidation treatment alone may not be satisfactory. The natural color of the BQ resulting from these high concentrations absorbs at the 420-mp wave length and will, therefore, interfere in the analysis. This problem can
HQ undergoes the same nitrosation reactions as M M H Q and will interfere
Table 111.
Recovery Deviation from Per cent average % -1.0 100.0 102.5 +1.5 101.8 $0.8 97.5 -3.5 105.0 +4.0 100.0 -1.0 102.5 $1.5 98.6 -2.4 Av. 101.0 +2.0
Determination of Known Concentrations of MMHQ in Acrylic Monomers
Monomer
MMHQ, P.P.M. Added Found
Ethyl acrylate Ethyl acrylate Ethyl acrylate
40
10
39
200
20 1
Butyl acrylate Butyl acrylate Butyl acrylate
16 200
15 47 20 1
2-Ethylhexyl acrylate 2-Ethylhexyl acrylate 2-Ethylhexyl acrylate Isodecyl acrylate
16 32 200 30
17 32 199 31
Acrylic acid Acrylic acid Acrylic acid Acrylonitrile
22 45 223 35
22 44 225 34
48
10
Recovery Deviation from Per cent average % 100.0 97.5 100.5
+0.4
93.8 97.9 100 5 106.3 100.0 99 5 103 3
-5.8 -1.7 +0.9 +6.9 +0.4 -0.1 +3.7
100.0
10 4 -2.0 +1.2 -2.5 =t2 1
97.6 100.8 97.1 Av. 99.6
-1.9 +0.9
0 8
0 6
"z Y
0 4
0, Q
0 2
0 0
0 WAVELENGTH m p
Figure 2. Absorbance-wave length relationship of nitrosated M M H Q in 2-ethylhexyl acrylate A. Sample vs. uninhibited acrylate blank 8. Blank VI. water
The precision and sensitivity of the M M H Q method, when applied to various acrylic monomers, are demonstrated in Table 111. These data were obtained with 1-cm. absorption cells, and under these conditions, 0.02 absorbance unit represents approximately 5 pg, of MhfHQ in the acrylates and acrylonitrile and 8 pg. in acrylic acid. The nitrosation method, as described for acrylic acid, is not applicable in the presence of HQ. LITERATURE CITED
(1) Afanas'ev, B. Ti'.,Khim. Prom. 1944, No. 7 , 18-19. 12) Critchfield. F. E.. Hutchison.' J. A . . ANAL.C H E ~32. . 862 (1960). (3) Ghosh, B., Bha'ttacharya, R., J . PTOC. Inst. Chemists (India) 18, 154-7 (1946). (4) Lacoste, R. J., Covington, J. R., Friscone, G. J., ASAL. CHEW 32, 990 (1960). (5) Stephen, W. J., Metallurgia 37, 333-4 (1948). (6) Stephen, IT. J., Belcher, R., Analyst 76, 45-59 (1951). RECEIVED for review October 31, 1960. Accepted January 30, 1961. \
be overcome by washing the sample, before analysis, with 5% sodium carbonate followed by two washes with 5% sodium bisulfite. The sodium carbonate oxidizes the H Q while the sodium bisulfite removes the resulting BQ as the water-soluble addition product. While the carbonate-sulfite treatment can be applied to ethyl acrylate
without loss of MMHQ, loss does occ-ir with the higher acrylates. For example, a loss of 17y0is incurred when the treatment is applied to butyl acrylate with a corresponding loss from 2ethylhexyl acrylate. However, since the higher acrylates normally contain no more than 0.02% HQ, the treatment should not be necessary for these monomers.
,
Fluorescent X-Ray Spectrometric Determination of Scandium in Ores and Related Materials ROBERT H. HEIDEL and VELMER A. FASSEL lnstitute for Atomic Research and Department o f Chemistry, Iowa State University, Ames, Iowa
b A fluorescent x-ray spectrometric method for the direct determination of scandium in ores, process materials, and rare-earth mixtures in the concentration range from 0.1 to 100 weight scandium oxide i s described. The samples are examined as blends of the powdered sample, vanadium pentoxide, and silicon carbide abrasive in the ratio of 5 : 3 : 3 . The intensity ratio of ScKa/VKa i s related to the concentration of scandium. A lithium fluoride crystal was used wiih a gas flow proportional counter as the detector. The coefficients of variation of the intensity ratios for a Soy0 and 1.5% scandium oxide standard were 1.3 and 3.7, respectively. Results obtained b y x-ray fluorescence and other methods are compared.
70
S
is one of the most widely distributed elements in the earth's crust ( I O ) . Hon-ever, it is the essential constituent of only one mineral, thortveitite. It is found in smaller quantities in Tviikite and bazzite. As much as 1% scandium niay be present in TI olfraniite and niiliute quantities are CAXDILX
found in eusenite, ccrite, orthite, thorianite, thorite, and other minerals (14). The scandium-containing minerals or ores and their process concentrates are usually of complex and widely varying composition. Therefore, the deterniination of scandium in these materials by gravimetric ( 1 1 ) , spectrophotometric (S),or titrimetric ( 4 ) methods requires extensive stcpwise separation of many, if not most, of the other constituents in the sample. Even with preliminary separations, interferences often arise from the remaining foreign substances. Optical emission spectrometric procedures have been used for the determination of scandiuni in specific matrices (5, 8.9, IZ), but Solodovnik, Rusanov, and Kondrashina (12) reported that for highest accuracy in determining scandium in substances of indeterminate composition, a preliminary separation of scandium and rare earths from the gross sample is likewise necessary. I n the present coniniunication i t is shown that the scandium content of minerals, ores, and products of their treatment can be determined directly by fluorescent x-ray spectroscopy with-
out any preliminary chemical processing or separations. EXPERIMENTAL
Fluorescent X-Ray Spectrum of Scandium. The scandium K a line a t
3.032 A. is observed with adequate sensitivity b y using a gas flow proportional counter and helium optical path. T h e inherent l o ~ vnoise of t h e detector and its discrimination against higher orders of scattered short ivave length radiation result in a signal-tonoise ratio favorable for minor constituent determinations. JVith a lithium fluoride crystal only the second-order erbium L & (3.028 h.)and tatitalum La1 (3.044 A , ) lines :ire sufficiently intense and close (within + 0.5' 2 8 ) t o the scandium K a linc to cause 110ssible spectral interfercncc. -it low scandium-erbium or scandiuni-tantaluin concentration ratios, the intprference from overlap, tliough small, is significant and it is nprcssary t o q7ply a correction. Both calcium and titanium niay be present along with scandiuni in natural and processed materials. Thcrefore, VOL. 33,
NO.
7,JUNE 1961
913
