Table I. Water on Tris(l,lO-Phenanthroline>Fe(II) Chelate
Blank area, mm2
1 . 6 5 x 10-3 5 . 0 0 x 10-3 10.00 x 10-3 13.20 X
4725 5340 5940 6054
4560 4560 4560 4560
5921 5977 6049 6076 6111
Water on Iron(III)-Benzohydroxamic Acid Chelate 95 0.0282 14.1 5826 131 0.0388 9.7 5826 223 0.0662 1 1 .o 5826 9.3 5826 250 0.0742 8.5 5826 285 0,0846
Corrected area, mmz
[HDI ( P I )
Dev from average
165 780 1380 1494
7 . 7 5 x 10-3 36.54 X 64.64 x 10-3 69.98 x
4.69 7.31 6.46 5,30
-1.25 +1.37 +0.52 -0.64
(Dev)* 1,5625 1.8769 0.2704 0.4096 4,1194 Total
u = *1.17
0.002 0.004 0.006 0.008 0.10
differential process. The large background water, due to the high solubility of water in the organic solvent, greatly limits the precision of the analysis. . Two identical columns were prepared and the flow rates were adjusted equally through both columns of the dual column instrument. Through one column was injected the chelate solution while the watersaturated solvent solution was injected through the other column simultaneously. The area of the resulting peak was due to water associated with the chelate. The differential method increased the precision of the analysis but it was difficult to establish equal flow rates. Geometry differences in the outlet block of the two channels resulted initially in “peaks” on both sides of the base line. This was due to the particular design of the detector used. Since the detecting element on the reference side was recessed, there was a time lag between its response and the sample side of the detector. In addition, the length of the column from the packing to the detector was longer for one column than the other. When an additional short piece of column was added to the short side, the peaks were normal again.
+3.6 -0.8 +0.5 -1.2 -2.0
12.96 0.64 0.25 1.44 4.00 19.29 Total
The differential technique is superior to the direct method because some solvents have large water solubilities and it is difficult to accurately detect small differences between relatively large numbers. 1-Decanol is a perfect example of this with its large water background and this is why it was used here to evaluate the method. The main disadvantage of the differential technique was the difficulty of establishing equal flow rates through both sides. However, once equal flows were established, many injections could be made without having to readjust the flow rates. The difficulty of preparing exactly identical columns was not as great as anticipated since a slight difference in column packing or length could be compensated for by a flow rate adjustment. RECEIVED for review August 26, 1966. Resubmitted April 7, 1970, and May 10, 1971. Accepted May 17, 1971. This research was supported by funds from the National Science Foundation and by the National Defense Education Act and this assistance is gratefully appreciated.
Gas Chromatographic Separation and Determination of Isomeric Methylbenzene Tricarbonylchromium Complexes Janet S. Keller,’ Hans Veening, and Bennett R. Willefoad Department of Chemistry, Bucknell University, Lewisburg, Pa. I7837
SINCEARENE TRICARBONYLCHROMIUM complexes were first synthesized in 1957 by Fischer and Ofele (I), much experimentation has been carried out involving synthesis, identification, and reactions of these compounds. Recent work in these laboratories has shown that gas chromatography (GC) can be used as a fast and accurate method for the separation and analytical determination of arene tricarbonylchromium complexes in mixtures (2). A combined mode of gas chromatography-mass spectrometry provides a convenient means for Present address, Department of Clinical Pediatrics, Hahnemann Hospital, Philadelphia, Pa. 1
(1) E. 0. Fischer and K. Ofele, Chem. Ber., 90, 2532 (1957). (2) H. Veening, N. J. Graver, D. B. Clark, and B. R. Willeford, ANAL.CHEM., 41, 1655 (1969). 1516
the identification of the separated compounds (3). G C has also been used to determine the cis-trans isomer ratios of alkylindane tricarbonylchromium complexes (4, 5), and the isomer distributions for the Friedel-Crafts acetylation of alkylbenzene tricarbonylchromium complexes (6). We report here the extension of the G C technique to include the separation and determination of ring isomeric methylbenzene tricarbonylchromium complexes. The compounds (3) W. J. A. VandenHeuvel, J. S. Keller, H. Veening, and B. R . Willeford, Anal. Lett., 3, 279 (1970). (4) D. E. F. Gracey, W. R. Jackson, W. B. Jennings, andT. R. B. Mitchell, J. Chem. SOC.( B ) , 1969, 1204. (5) D. E. F. Gracey, W. R. Jackson, C. H. McMullen, and N. Thompson, ibid., p 1197. (6) W. R. Jackson and W. B. Jennings, ibid., p 1221.
ANALYTICAL CHEMISTRY, VOL. 43, NO. 11, SEPTEMBER 1971
studied include the isomeric di-, tri-, and tetra-methylbenzene tricarbonylchromium complexes. EXPERIMENTAL 3-
Samples. The following complexes (I-IX) were prepared by methods similar to those described in the literature (7, 8).
I 11 111 IV
1,2-dimethylbenzenetriicarbonylchromium(o-XTC) 1,3-dirnethylbenzenetricarbonylchromium(m-XTC) 1,4-dirnethylbenzenetricarbonylchromium (p-XTC) 1,2,3-trimethylbenzenetricarbonylchromium (1,2,3-
TBTC) V VI VI1 VI11 IX
1,2,4-trirnethylbenzenetricarbonylchromium(1,2,4TBTC) 1,3,5-trirnethylbenzenetricarbonylchromium(1,3,5TBTC) 1,2,3,4-tetramethylbenzenetricarbonylchromium (1,2,3,4-TMTC) 1,2,3,5-tetramethylbenzenetricarbonylchromium (1,2,3,5-TMTC) 1,2,4,5-tetramethylbenzenetricarbonylchromium (1,2,4,5-TMTC)
In general, the compounds are synthesized by refluxing chromium hexacarbonyl with the deaerated ligand in an inert solvent such as diglyme under nitrogen. The complexes are removed by precipitation and filtration and are purified by vacuum sublimation. Each of the complexes was characterized by melting point, infrared and NMR spectra, and elemental analysis. Apparatus and Instrumental Conditions. A Perkin-Elmer Model 900 gas chromatograph equipped with a flame ionization detector was used to separate and detect the complexes. A Leeds and Northrup (1 mV) recorder with a disc integrator was used to record and integrate the chromatograms. Helium was used as a carrier gas, and prepurified hydrogen and air were used to operate the flame ionization detector. Two types of columns were employed for this work. One of these was a 6-ft (or 12-ft) length of borosilicate glass tubing, 2 mm i d . , packed with 100-120 mesh Gas Chrom-Q coated with 3 . 6 z SE-30; the second column consisted of a 100-ft X 0.5-mm i.d. stainless steel, support coated open tubular (SCOT) column coated with rn-bis(rn-phenoxyphenoxy) benzene and Apiezon L (Perkin-Elmer). The injection block for the SCOT column was equipped with a split ratio restrictor of 1:4. The hydrogen flow rate was 24 ml/min while that of air was 300 ml/min. Other conditions of operation are given in the tables and figures. Procedure. Solutions of the complexes were prepared in spectral grade benzene or carbon tetrachloride solvents, which were deaerated with prepurified nitrogen before use in order to minimize oxidation of the chromium complexes. The flow rate for the SCOT column was determined by injecting 40 to 50 fil of methane and measuring the time of elution. The resultant linear gas velocity in ft/sec was then converted to volume flow rate in ml/min by the appropriate conversion factors. Flow rates for the packed column were determined by use of a soap-film flow meter. For quantitative studies, the internal standard method described previously (2) was employed, except that peak areas were measured by means of a disc integrator. Chromium complexes which did not interfere with the determination under study were used as internal standards in the calibration solutions and in the synthetic mixtures. Calibration graphs used for the analysis of synthetic mixtures were constructed by drawing a computer calculated least squares straight line through the experimental points obtained by plotting ratios of peak areas (sample to internal standard) us. concentration of metal complex (g/ml). (7) B. Nicholls and M. C. Whiting, ibid., 1959, 551. (8) E. 0. Fischer, K. Ofele, H. Essler, W. Frohlich, J. P. Mortensen, and W. Semmlinger, Chern. Ber., 91, 2763 (1958).
Time ( m i d
Figure 1. Separation of 1,3,5-TBTC, 1,2,4-TBTC, and 1,2,3-TBTC on a packed column Internal standards: p-XTC and 1,2,3,4-TMTC Column temp: 80 to 200 "C, programmed at 4"/min Inj. port temp: 145 "C Column flow: 11 ml/min Amounts injected: 1.49 pg 1,3,5-TBTC 1.68 ,up 1,2,CTBTC 2.09 pg 1,2,3-TBTC
Table I. Determination of Isomeric Trimethylbenzenetricarbonylchromium Complexes in Mixtures Column, 6-ft packed; column temp, programmed from 80 to 200 "C; programming rate, 4"/min; carrier gas, helium; column flow rate, 11 ml/min; inj. port temp, 145 "C; detector temp, 150 "C; internal standard, 1,2,3,4-TMTC Sample, giml X lo4 Error, % Complex Taken Found 1,3,5-TBTC 2.02 2.20 +8.9 3.03 3.48 +14.8 5.05 5.25 +4.0 10.04 10.20 +1.6 12.37 13.63 +10.2 -3.5 2.22 1,2,4-TBTC 2.30 3.85 +11.3 3.46 5.76 5.97 +3.6 9.92 10.00 -0.8 13.95 14.00 -0.4 2.71 +8.0 1,2,3-TBTC 2.51 4.17 3.77 +10.6 6.38 +1.6 6.28 10.47 -2.1 10.70 16.96 -2.9 17,46 Average error = =t5.6%.
RESULTS AND DISCUSSION It was found that the isomeric trimethylbenzene complexes, 1,2,3-TBTC, 1,2,4-TBTC, and 1,3,5-TBTC could be separated and determined on the packed column as shown in Figure 1. 1,2,3,4-TMTC and p-XTC were included in this mixture and served as the internal standards. Identification of the peaks was accomplished by trapping the eluted components in hexane and recording a UV spectrum of the resulting solution. In each case, the UV spectrum of the eluted sample was identical with that of the known compound. Quantitative results obtained for the determination of the isomeric trimethylbenzene complexes o n a packed column in mixtures are given in Table I. In each case, the calibration plots used for the analysis of synthetic mixtures were found to be linear g/ml. The accuracy of between 2.0 X IO-' and 2.0 X
ANALYTICAL CHEMISTRY, VOL. 43, NO. 11, SEPTEMBER 1971
Figure 2. Gas chromatogram of the peaks resulting from the injection of the di-, tri-, and tetramethylbenzenetricarbonylchromium complexes on a SCOT column Column temp: 80 "C for 2 min. grammed to 150 "C at 4"/min Inj. port temp: 180 "C Column flow: 6.2 ml/min Complex p-XTC m-XTC 0-XTC 1,3,5-TBTC 1,2,4-TBTC 1,2,3-TBTC 192,4,5-TMTC 1,2,3,5-TMTC 1,2,3,4-TMTC
Amount injected, pg 2.23 2.06 2.07 2.06 1.76 2.76 3.20 1.84 1.69 L
the determination is good to within * 5 . 6 % relative error. It should be noted that synthetic mixtures containing ca. 3.5 X 10-4 g/ml of complex gave consistently high positive errors; this causes a relatively high overall error. Omission of these data results in an over relative error of *4.0%, more nearly comparable to results obtained in other experiments. While arene tricarbonylchromium complexes survive gas chromatographic analysis intact on a packed glass column, elution of these compounds on the SCOT column resulted in
complete decomposition. The peaks resulting from complex pyrolysis were those of the hydrocarbon ligands. A temperature programmed chromatogram showing the peaks for the eluted ligands resulting from decomposition of a nine-component mixture of all the complexes in CC14 solution on the SCOT column is shown in Figure 2. An additional peak at 15.8 minutes was identified as Cr(C0)o. The chromatogram obtained when the complexes are injected is identical to that obtained when a mixture of the free ligands is eluted under
Pyrolytic Determination of Isomeric Di-, Tri-, and TetramethylbenzenetricarbonylchromiumComplexes in Mixtures Trimethylbenzenesc Tetramethylbenzene+ Dimethylbenzenes* Sample, g/ml x lo4 Sample, g/ml X lo4 Sample, g/ml X lo4 Complex Taken Found Error, Complex Taken Found Error, Complex Taken Found Error, % 2.35 -3.4 +1.7 1.85 2.03 1.96 2.31 -5.6 1,3,5-TBTC 1,2,4,5-TMTC p-XTC 1.96 3.08 3.47 -3.4 3.56 $1.3 +2.6 2.85 3.04 2.95 3.74 4.63 -1.9 +5.4 2.89 4.06 -2.0 4.88 2.95 -0.6 +2.6 3.51 5.07 5.04 5.79 -9.2 5.94 3.93 8.11 7.11 5.19 5.68 -4.2 -1.9 1.86 1.66 -2.5 12.15 9.50 +5.1 +2.6 11.56 9.26 10.00 9.80 -2.0 -1.3 12.11 11.95 -5.4 2.08 2.03 -2.4 m-XTC 2.02 1.91 1,2,4-TBTC 2.14 2.31 2.45 +6.1 3.03 1,2,3,5-TMTC -1.3 2.08 +2.9 2.99 3.12 2.96 -5.1 2.31 2.48 $7.4 3.03 +0.3 3.04 2.19 -5.2 -5.5 5.19 5.33 $2.7 2.31 4.17 5.05 8.31 1.68 -7.6 3.46 3.52 +1.7 -4.0 8.08 1.76 -2.2 13.24 14.03 5.17 5.64 10.65 +6.0 -0.8 10.57 9.22 +0.2 9.24 2.01 2.10 +4.5 15.09 14.96 +0.1 2.04 1.87 -8.3 1,2,3-TBTC o-XTC 2.01 2.06 +2.5 2.04 2.03 -0.5 2.61 3.01 -5.6 1,2,3,4-TMTC 2.60 +0.4 3.05 3.04 +1 .o 2.88 4.02 2.60 +0.8 4.01 -0.2 2.62 4.24 +4.4 4.06 +1.5 3.89 7.24 -10.0 3.95 8.04 8.13 7.85 -3.4 -3.1 5.19 13.70 14.84 5.03 11.10 11.19 +8.3 +0.8 9.91 -4.5 10.38 14.99 +1.1 14.83 Average error = +3.6 Average error = * 4 , 2 z Average error = & 2.6 Z a SCOT column, helium carrier gas. * Column temp, 80 'C; Row rate, 3.9 ml/min; inj. port temp, 180 "C; detector temp, 210 "C; internal standard, toluenetricarbonylchromium. c Column temp, 120 "C; Row rate, 4.1 ml/min; inj. port temp, 190 OC; detector temp, 192 'C; internal standard, p-XTC. d Column temp, 115 "C; Row rate, 5.8 ml/min; inj. port temp, 170 "C; detector temp, 180 "C;internal standard, 0-XTC. Table 11..
ANALYTICAL CHEMISTRY, VOL. 43, NO. 11, SEPTEMBER 1971
Figure 3. Partial separation of the di-, tri-, and tetramethylbenzenetricarbonylchromium complexes on a 12-foot packed column Column temp: 80-180 "C at 6"/mniir Inj. temp: 145 "C Column flow: 10 ml/min Amounts injected: approximately 1 pg of each complex Table 111. Relative Retention Data SCOT column temp. programmed Packed column (hydrocarbon (SE-30, 12-ft) peaks from temp. programmed pyrolyzed complexes) (eluted complexes) Re1 Re1 retention retention Ret time, (p-XTC Ret time, (p-XTC Compound min = 1.00) min = 1.00)
p-xi'c: m-XTC 0-XTC l13,5-TBTC 1,2,4-TBTC 1 ,2,3-TRTC 1,2,4,5-TMTC 1,2,3,5-TMTC 1,2,3,4-TMTC
6.6 6.8 7.8 10.7 11.7 13.4 17.2 17.6 19.2
1.00 1.03 1.18 1.62 1.77 2.02 2.60 2.66 2.91
34.8 35.4 37.2 38.1 39.6 43.2 47 2 48.7 52.5
1 .OO 1.02 1.07 1.09 1.14 1.24 1.35 1.40 1.51
identical conditions, with the sole exception of the appearance of the Cr(CO), peak. Though the observed peaks are those of pyrolysis products, the quantitative determination of these complexes in mixtures has been shown to be feasible. Table I1 shows the quantitative results obtained for the analysis of synthetic mixtures of the di-, tri-, and tetramethylbenzenetricarbonylchromiumcomplexes. Toluenetricarbonylchromium (TTC), p-XTC, and o-XTC served as internal standards for these three groups of compounds. The average errors found in these three analyses were 1 3 . 6 , 1 4 . 2 , and 1 2 . 6 % , respectively; this indicates that pyrolysis was quantitative when using the SCOT column. It is likely that the decomposition of these compounds occurred in the open tubular column injection accessory rather than in the column. Contact of the gaseous complexes with heated stainless steel is prolonged in the injection block. Decomposition on the column is thought to be unlikely because the retention times were reproducible under a given set of
conditions, and because the column was internally coated with powdered quartz and stationary phase. While it was possible partially to separate and elute all the complexes without decomposition o n a packed glass column, the resolution for the di- and tetramethylbenzene complexes was not sufficient to warrant a quantitative study. The elution order for the nine alkylbenzenetricarbonylchromium complexes studied was p-XTC < m-XTC < 0-XTC < 1,3,5TBTC < 1,2,4-TBTC < 1,2,3-TBTC < 1,2,4,5-TMTC < 1,2,3,5-TMTC < 1,2,3,4-TMTC. A temperature-programmed chromatogram of these complexes o n a 12-foot packed column is shown in Figure 3. It is interesting to note that, for each group of isomers, the most symmetrically substituted complex elutes first, whereas the vicinally substituted complex elutes last. The retention times and relative retentions of all the complexes on the packed column compared to those found for the pyrolysis products on the SCOT column are shown in Table 111. It is worth noting that the differences in relative retentions for isomers within each group are far more pronounced for the SCOT column than for the packed column, indicating the superior resolution obtained on the former. Also, the pyrolysis technique is much faster. The analytical determination of isomeric methylbenzenetricarbonylchromium complexes by their elution from a packed column, and by measurement of the eluted hydrocarbon peaks obtained from metal complex injections o n a capillary column has been found to be a fast, accurate, and reliable method. Satisfactory analytical results obtained, whether or not the complexes retain their integrity, have shown that the use of GC has greatly facilitated the determination of metal *-complexes of this type. RECEIVED for review March 19, 1971. Accepted June 2, 1971. Paper taken in part from the M.S. thesis of J.S.K. Presented at the 161st National Meeting, American Chemical Society, Los Angeles, Calif., March 1971. Work supported by the National Science Foundation Grants GP-8938 and GP-18755.
ANALYTICAL CHEMISTRY, VOL. 43, NO. 11, SEPTEMBER 1971