High gas temperature furnace for species ... - ACS Publications

0. 6 l6. 24. Figure 4. HPLC of (A) blank of field soil and (B) field soilspiked TU at the level of 160 ng/mL. (1) TU, (2) DMTU as internal standard. C...
0 downloads 0 Views 662KB Size
ANALYTICAL CHEMISTRY, VOL. 51, NO. 3, MARCH 1 9 7 9

387

applied to industrial waste, river water, seawater, and secondary treated sewage effluent, and none of the samples showed interfering peaks; T U was not detected.

ACKNOWLEDGMENT T h e author thanks T. Akiyama, Research Director of the Kitakyushu Municipal Institute of Environmental Health Sciences, for his valuable advice and encouragement. He also thanks the authority of the institute, for permission to publish this paper.

2

LITERATURE CITED

1 9

It

2q

1

0

1

iieteition t i r e

-

I E

I'

Tin

i

I6

H. J. Morris, J . Nafl. Cancer Inst., 7, 159 (1946). H. D. Purves and F. W. Grissbach, Br. J . Exp. Pafhol., 27, 24 (1946). H. D. Purves, B r . J . Exp. Patbol., 27, 294 (1946). A. Rosin and M. Bachmilewitz, Cancer Res., 14, 494 (1954). M. Watanabe, Hiryo Nenkan (in Japanese), 1978, p 337. R. C. Hoseney and K . F. Finny, Anal. Chem., 36, 2145 (1964). Association of Official Analytical Chemists, "Official Methods of Analysis", 11th ed., Washington D.C., 1970, p 350, 29.099, E. T. Raktzes, Anal. Chim. Acta, 78, 495 (1975). S. Goto, 0. Ogawa, I . Asakawa, and C. Oshima, Nippon Kogyo Kaishi, 88, 1067 (1972). W. P. Mckinley and R. Yasin, J. Assoc. Off. Anal. Chem., 43, 829 (1960). J. H. Onley and G. Yip, J . Assoc. Off. Anal. Chem., 54, 165 (1971). M. B. Devani. C. J. Shishoo, and B. K. Dadia, J . Chromatogr.. 105, 186 (1975). I . M. Kovina and N. K . Rozhkova, Dokl. Akad. Nauk. Uzb. SSR, 25, 25 (1968). K. Seiffarth and W. Aldelt, Plaste Kautsch., 15, 818 (1968). E. N. Boitsov, Yu. I. Mushkin, and V. M. Kavlik, Zavod. Lab., 35, 790 (1969). N. Nornura, D. Shiho, K. Osuga, and M. Yarnato, J . Chromatogr., 42, 226 (1969). R. N. Sargent and W. Riernan 111, J . Phys. Chem., 61, 354 (1957). K. Fujita, Y . Arikawa. and S. Ganno, Nippon Kagaku Kaishi, 3, 463 (1975). J. L. Waters, J. N. Little, and D. F. Horgan, J , Chromatogr. Scl., 7, 293 (1969). L. R. Snyder. J . Chromatogr. Sci,, 7, 352 (1969).

24

Figure 4. HPLC of (A) blank of field soil and (B) field soil spiked TU at the level of 160 ng/mL. (1) TU, (2) DMTU as internal standard. Conditions as in Figure 3. Sample volume injected: 1 0 0 pL, U V monitored: 2 3 7 nm, AUFS: 0.08

T h e limits of detection were estimated to be 1.5 ng, 8.8 ng, 2.3 ng, 3.7 ng, 3.7 ng, and 9.1 ng for T U , TAA, E T U , DMTU, ATU, and D E T U , for a standard solution, with 100-pL inject ions, respectively. Analysis of TU Spiked to Field Soil Sample. Typical chromatograms of field soil samples are shown in Figure 4A and 4B. Control field soil samples showed no interfering peaks (Figure 4A). A recovery of 99.3 f 2.66% (mean f rel. stand. dev., n = 5) was obtained when T U was spiked to field soil samples at the level of 160 ng/mL. The limits of detection were estimated to be 2.7 ng in this case. The method was also

RECED-ED for review March 25, 1977. Resubmitted September 7. 1978. Accepted November 8, 1978. This work was financially supported by the Environment Agency of Japan.

High Gas Temperature Furnace for Species Determination of Organometallic Compounds with a High Pressure Liquid Chromatograph and a Zeeman Atomic Absorption Spectrometer Hideaki Koizumi, * ' Ralph D. McLaughlin, and Tetsuo Hadeishi Lawrence Berkeley Laboratory, University of California, Berkeley, California 94720

Species determination of organometallic compounds is nowadays one of the most important subjects in analytical chemistry. This is because the toxicity of metals depends upon the binding conditions in the compound ( I ) . Speciation of organometallic compounds is also important in physiology and biology. Up to now, more than 1000 enzymes and coenzymes have been found a n d about one third of them are "metalloenzymes" or "metal-substrate complexes" (2). It is a well known fact t h a t commercial gasoline contains alkyllead compounds. In the United States, leaded gasoline contains around 0.19'0 of alkyllead, and commercial unleaded gasoline is defined as gasoline having not more than 0.05 g Pb/gal (13.16 pg/mL) ( 3 ) This paper describes the use of a special furnace Zeeman atomic absorption (ZAA) combination as a detector for a high pressure liquid chromatograph (HPLC) for the determination of organometallic lead com-

A new furnace has been constructed that allows atomic absorption detection of volatile organometallic compounds. The operation of this furnace is demonstrated by analyzing the eluent of a high pressure liquid chromatograph utilizing Zeeman atomic absorption spectrometry. The content of tetraethyllead in National Bureau of Standards gasoline standards was determined. Data are presented on the ability of this furnace to suppress interference with cadmium and lead determinations by MgCI,, CuCI,, and CaCI,. I t was found that two orders of magnitude more interferent can be tolerated. The determination of lead in automotive exhaust is also described.

'Permanent address, Naka Works, Hitachi Ltd., Katsuta, Ibaraki

312, Japan.

0003-2700/79/035 1-0387$01 O O / O

C

1979 American Chemical Society

388

A N A L Y T I C A L CHEMISTRY, VOL. 51, NO. 3, MARCH 1979

pounds in gasoline. Various types of organic solvents are used as eluents for H P L C columns: aromatic compounds such as benzene and toluene, alcohols, ketones, and nitro compounds. Therefore, we are obliged to determine trace amounts of metals in a variety of organic solvents. With conventional techniques of trace element detection, including atomic emission and absorption spectroscopy and mass spectroscopy, i t is very difficult t o avoid interferences caused by large amounts of organic solvent. I n many cases, the organic solvent cannot be eliminated by vaporization before the measurement because many organometallic compounds have boiling points below 200 "C and a considerable part of the organometallic compound escapes with vaporization of the solvent. For example, the mercury signal from methylmercuric chloride in benzene solution cannot be obtained by mass spectrometry, because this compound evaporates a t around 100 "C ( 4 ) . Several researchers have used atomic absorption methods as detectors for H P L C (5-7). These methods generally have many advantages over other techniques. However, interference caused by the eluent was still a problem, because ashing temperatures had to be kept as low as possible to avoid evaporation of organometallic compounds. Strictly speaking, furnace ashing should not be used for organometallic compounds because of their volatile nature. T h e Zeeman atomic absorption (ZAA) technique has features that make it attractive for determinations of this kind (8). ZAA can measure accurate values of atomic absorption signals with 500-1000 times larger background absorption ( S I I ) . The stability of the base line is also a desirable feature for long time measurement with H P L C (12). However, a t times, there is a problem in the determination of organometallic compounds, even though the ZAA technique provides excellent background correction. For example, the absorption signal of P b in alkyllead compounds cannot be observed with the conventional graphite furnace because highly volatile alkyllead escapes from the cuvette before atomization. There are also problems t h a t arise when the flame is used for atomization of alkyllead in gasoline which require special treatment of the sample to obtain accurate results. These problems arise because tetramethyllead (TML) produces a different absorbance value than does the same amount of tetraethyllead (TEL) (13). It is much more difficult to get the same absorbance response from the various species in conventional furnace AA than in flame AA. Therefore, a different furnace is necessary to atomize the metal in organometallic compounds and for speciation studies in conjunction with HPLC. One way to minimize the different response of different species is to dissociate the organometallic compound as completely as possible into individual atoms by heating to high temperatures. Conventional furnaces are not very effective because the gas temperature of the furnace is much lower than the wall temperature and the sample vapor easily diffuses out of the hot part of t h e cuvette before reaching a high temperature (14, 15). Figure 1 shows various possible furnace designs to achieve high gas temperatures. Figure l a shows the type of furnace used in this work. Figure I b shows a double chamber furnace ( I O ) cuvette which can be used in commercially available Massmann type furnaces. T h e same high temperature can be obtained with this arrangement as with the conventional graphite tube cuvettes because the surface area of the cuvette is kept small to minimize radiation loss. T h e recovery of the element in a sample with a complex matrix is highly improved compared with conventional cuvettes. However, this cuvette cannot achieve 100% recovery for highly volatile organometallic compounds such as T M L and TEL. Figure ICshows

Inner chomber

Graphite tube

(b)

Light c _

k

I O mm

(C)

Figure 1. Various types of high gas temperature (HGT) furnaces

Table I. Instrumentation for HPLC-ZAA System HPLC high performance liquid chromatograph, Hitachi M633, 0-350 kgicm', 0.36 - 3.6 mL/min column, Hitachi, 2 . 5 X 500 mm resin, Hitachi Gel N o . 3010 Furnace graphite, Ultra Carbon 0.5-in. diameter porous graphite, RVC 1 0 0 PPI porosity grade furnace power supply, reactor controlled, 20 V, 700 A ZAA spectrophotometer light source, magnetically confined lamp (dc + rf ( 5 0 MHz)) magnet, permanent, 1 2 kg variable retardation plate, 0 - h / 2 , 30 Hz polarizer, Rochon prism (quartz, optical contact) monochromator, Hitachi MlOO spectrophotometer photomultiplier, Hamamatsu T.V., YA 7 1 2 2 chopper, Bulova L2C, 1.0 KHz electronics lock-in amplifier (including log convertor, AGC) recorder, Honeywell Electronik 1 7 a closed system furnace. A sample is atomized in a closed space and kept there for a while to achieve complete dissociation. The piston is moved with the thermal expansion of the sample vapor, and just after the piston passes by the windows, an absorption measurement is made. The properties of this cuvette are now being investigated. The circular cavity furnace (161,which was recently developed by L'vov. is also a kind of high gas temperature furnace. If a furnace could be constructed to raise the gas temperature high enough, the dissociation of all organometallic compounds would occur and the signal would be independent of species. In this paper, we will report the species determination of organometallic lead compounds by utilizing a HPLC-ZAA system with a new furnace t h a t can efficiently atomize these highly volatile compounds.

EXPERIMENTAL Table I lists the experimental apparatus and parameters. The same ZAA spectrophotometer as reported before was used through this experiment ( 1 1 ) . Figure 2 shows the cross section of the present furnace, which is called the high gas temperature (HGT) furnace. This furnace consists of several separate parts; the sample cup, the thermal converter and reactor of porous graphite, a narrow hole (which at high temperatures also acts as a thermal converter), and the absorption cell. The sample vapor flows through the thermal converter and its temperature is raised sufficiently to decompose the compound and to atomize the metal. After that, the sample vapor is carried t o the absorption cell for ZAA measurement.

ANALYTICAL CHEMISTRY, VOL. 51, NO. 3, MARCH 1979

389

m

Graphite

Is3

Porous graphite

m

EEa

Brass

Absor'P 1.ight beam

Figure 2. Cross section of HGT furnace Table 11. Properties of Selected Pb Compounds ( 2 0 ) 312'

critical temperatures, C compound

_.

_. _~

~~

>;; CJrrer'

~-

~~~

>-,__

i

.

A!

Figure 3. Relationship between current and temperature of hottest region of HGT furnace In the case of the present furnace, the center portion of the cuvette is heated to a high temperature first because of the small heat capacity and the large electrical resistance. Conductive heating causes the porous graphite to be heated next and, finally, the tantalum cup is heated. A few seconds after the current is turned on, the temperature difference between the three sections begins to decrease. Vaporized sample in the cup follows the flowing argon gas through the porous graphite where its temperature is raised by coming into intimate contact. It then passes through a small hole, the walls of which have the highest temperature. Because of intimate contact with these surfaces, the gas temperature becomes equal to the wall temperature before passing into the absorption cell. Figure 3 shows the relationship between the temperature of the regions just above the absorption cell and the current supplied to the cuvette. The surface area and the heat capacity of this cuvette are as small as the Massmann furnace (17). It is easy to raise the temperature of the cuvette to 2800 "C by supplying about 5 kW of electric power. This high temperature makes it possible t o atomize various types of elements that are not possible with the Woodriff type furnace (18, 29). This furnace is small enough

lead nitrate lead chloride lead monoxide lead chloride tetramethyllead tetraethyllead tetraphenyllead a

formula

melting boiling dissociation

Pb(NO,), PbCI, 501 950 PbO 888 1740 PbC1, 501 950 Pb(CH,), - 27.5 110 Pb(C,H,), - 136.S 200a Pb(C,H,), 227.7

470

200 270

Decomposition temperature

to fit between the poles of a permanent magnet if the polarized Zeeman AA technique is utilized (12). The system used to separate alkyllead compounds is described in Table I. Methyl alcohol was used as the eluent. The pressure was about 30 kg/cm2, and the flow rate was 0.67 mL/min. A sample of 10 pL was injected into the HPLC while the flow was stopped. A 10-pL aliquot from each 250-pL portion of column effluent was intermittently introduced into the HGT furnace.

RESULTS A N D D I S C U S S I O N Table I1 shows the properties of lead and some lead compounds (20). Usually, P b ( N 0 J 2 is used for t h e lead standard in atomic absorption spectrometry. P b ( N 0 J 2 decomposes into P b O around 470 "C, and atomic P b is produced from the dissociation of PbO. In t h e graphite furnace, the following reduction process also proceeds on the surface of graphite. C PbO = P b CO

+

+

Atomic lead appears at around 770 "C, and the change of free energy. JG, for this reaction is --22 kcal/mol at this tem-

390

ANALYTICAL CHEMISTRY, VOL. 51, NO. 3, MARCH 1979 Pb (C,H,

)4

5

f r o m HPLC ~

bi

4

3 2 Time (a) ZAA signals of tetraethyllead peak from HPLC column showing distribution in 1.5 mL of eluent, (b) signal from standard lead solution (0.5 N "0,) Figure 4.

1

0 5

Table 111. Results of Direct DeterminationC of Pb in NBS Gasoline Standard by ZAA with High Gas Temperature Furnace Pb in NBS SRM 1636, pg/mL'

standard used Pb(NO,), Pb(C,H,), At 20 "C. nations, 6.

solvent

resultb

certified value

HNO, [ 0 . 5 N ] 1 9 . 1 t 0.5 C,H, 19.3 + 0.5 "," Solvent: MIBK. Number of determi-

perature. Usually, the ashing temperature for P b is chosen to be below 600 "C, and the atomizing temperature is around 1500 "C (21). TML and TEL are liquid at room temperature. TML boils a t 110 "C and T E L decomposes a t 200 "C. They are easily vaporized, even a t the drying temperature. With the conventional furnace, if we try to atomize these organometallic compounds without drying, they vaporize before the gas temperature becomes high enough to produce lead atoms. T h e organometallic lead compounds listed in Table I1 were introduced into the present furnace and the signals were compared with those from an inorganic lead standard: P b ( N 0 3 ) 20.1 pg/mL in HNO, (0.5 N ) . The alkyllead compounds were dissolved in benzene, such t h a t the final concentration was 0.1 pg/mL; 1O-pL aliquots were directly atomized without drying and ashing. The strong background absorption interference was corrected by using ZAA technique. T h e same peak height of lead atomic absorption was observed for each sample. Examples of the signal shape for organic and inorganic lead compounds are presented in Figure 4. T h e similarity in peak shape supports the idea t h a t the metal organic compound is totally atomized in the furnace. T h e concentration of P b was determined in the NBS standard fuel SRM-1636 which contains the lead in the TML and TEL forms in 91 octane number reference fuel (a mixture of about 91 YC 2,2,4-trimethylpentane and 9% n-heptane). This N B S standard was diluted by a factor of 1/200 with methyl isobutyl ketone, and 10 pL of the sample was directly atomized using the H G T furnace. Both inorganic and organic lead solutions were employed as standards: Pb(YO3), in HNO, (0.5 N), and tetraphenyllead ( T P L ) in benzene. T h e reason T P L was used as the organic lead standard is that T P L is crystalline a t room temperature and is comparatively stable. This aids in the preparation of an accurate standard solution. Table I11 shows the results of the analysis. Good agreement was obtained with the certified NBS value, both when the inorganic and organic lead standard solutions were used. From these results, we can safely conclude t h a t both the same accuracy and the same sensitivity is obtained for the inorganic lead and organic lead compounds in spite of a big difference

-

I

I

I

I

4

I

SRM 1636-A Pb:

F 3

v

23

a

2 1

C 5 4

SRM 1636-6 ( P b :535p.g/mL )

3 X l w

2

IJ

1

--XiG

1 I

C

2

4

6

1

8

1

I

1

I

1

0

3

6

9

12

10 Volume(mL) I

15 Time ( m i n )

Figure 5. Elution chromatograms for alkyllead compounds: (a) 80 pg/mL tetramethyllead and 20 WglrnL tetraethyllead, (b) and (c) NBS

standard lead in gasoline samples in vaporization temperatures. Now, the determination of the concentration of various organometallic species is possible with the H G T furnace because the sensitivity is independent of species. Figure 5a shows the histogram of lead concentration in various portions of the eluent. The peaks of T M L and T E L appear at retention times of 5.5 and 10 min, respectively. The histogram in Figure 5a shows t h a t T M L and T E L can be separated without overlap. T h e recovery ( t h e ratio of the amount of injected lead to the total amount of the observed lead in the elutent from the column) was 97% for T E L , and 25% for TML; 75% of T M L might be dissociated and trapped in the resin. In the GC-Flame AA system, several T M L peaks

ANALYTICAL CHEMISTRY, VOL. 51, NO. 3, MARCH 1979

391

Table IV. Comparison of Chemical Interferent Levels between High Gas Temperature Furnace and Conventional Furnaces recovery, % concentration of interferent,

element Pb

interferent MgC1,

pg/mL 100 1000 5 000 10 0 0 0 100

CaCl,

1000 5 000 10 0 0 0

CUCl, Cd

100 10 000 1 0 000

CaCl,

conventional furnace

ref.b

6 3 (95)' 6 (29)' 18

25, 26 25 27

74 (98)' 0 (45)' 48

25 25 23

high gas temperature furnace

93 68

100 0

24

0

23

82 97

' CTA. Ref. 2 3 : HGA70, HGA2000. Ref. 24: CRA63. Ref. 2 5 : CRA63 (a.CTA). Ref. 26: HGA.70. Ref 27: HGA2200.

" -

--+ - 6

9

IC

2

"Cl"rnt

m_

2

15

8

- r e

rir

Flgure 6. Chromatogram showing spreading of Pb(NO,), by the Hitachi Gel No. 3010 column

were observed (22). This shows that T M L is very unstable. When t h e inorganic lead solution P b ( N 0 3 ) , was introduced into the H P L C containing Hitachi No. 3010 resin, almost no lead was recovered (Figure 6). In order to separate organic lead from inorganic lead, the Hitachi ion-exchange column No. 2611 should be used. We have already reported on the use of this column to separate Co bound to vitamin B,, from inorganic Co ( 8 ) . Next, the determination of T M L and T E L in the NBS standard was performed using the HPLC-ZAA system. Ten p L of NBS 1636-A (10.15 Kg P b / m L ) was introduced into the system. A clear peak (Figure 5b) was obtained at the retention time a t 10 min. The concentration of lead in the elutent is also shown in this figure. The content of lead integrated over this peak is 187 ng, which is 97.6% of the lead in the injected N B S standard (191.5 ng). No peak could be observed a t the retention time corresponding to T M L a t 5.5 min. A higher concentration standard ( N B S 1636-B, 535 pg P b / m L ) was introduced into the system to check the existence of T M L . However, a peak for T M L could not be observed as shown in Figure 5c. We also confirmed that inorganic lead and T P L do not cause a peak near the T E L position (Figure 6). Therefore, we can conclude t h a t the N B S standards which were used in this experiment contain more than 95% of one lead species (TEL). (These N B S standard P b in gasoline samples were purchased in 1975 and the unopened ampules were stored in a dark box until used.) D e c r e a s e of Chemical I n t e r f e r e n c e s Caused by L a r g e Q u a n t i t i e s of Salts. Chemical interference by metal halides causes serious problems in conventional furnace AA (23-30). T h e use of the high gas temperature (HGT) furnace greatly reduces this type of interference. A typical case of this kind of interference occurs when MgC12 is present during t h e determination of P b . Sometimes only 50 wg/mL of MgCI, causes a 30% depression of the P b signal. Molecular absorption and scattering by MgCl, is not large a t 283.3 nm (the

-

I

-d

s

3oicc

I

Time

Figure 7. Lessening of suppressions of Pb signal by MgCI, through the HGT furnace

the use of

wavelength of the P b resonance line). Therefore, this phenomenon is totally due to chemical interference. Figure 7 shows the effect of MgC1, upon the P b signal using the H G T furnace. Table IV lists the magnitude of signal suppression for several salts. I t can be seen that MgC12 is the strongest suppressing agent. This ability of the H G T furnace to prevent signal suppression cannot be explained by the high temperature of the sample vapor. T h e change of free energy, AG, for the dissociation of PbCl is still large a t a temperature of 2500 "C. Hence, thermal dissociation would not be expected to occur. A possible explanation would be t h a t a reduction process is accelerated because of the large surface area of the porous graphite. O t h e r A p p l i c a t i o n s of t h e H i g h G a s T e m p e r a t u r e F u r n a c e . T h e lead concentration in exhaust gases can be measured by using the H G T furnace. Automobile exhaust gas was collected in a polyethylene bag and was forced to flow through the H G T furnace carrier gas inlet port a t a flow rate of 90 cm3/min. The lead concentration in the exhaust gas was determined from the area under the absorption signal to be 0.3 ng/cm3. Hence, this furnace is capable of on-line monitoring of trace elements in gas or ambient air. Another important area of application is to the direct analysis of solid samples. Because this furnace tends t o prevent molecular formation, solids can be volatized and the elements in t h a t sample can be atomized, and the sensitivity will be largely independent of organometallic species. This will be true, even for highly volatile compounds. Preliminary work on metal organic lead compounds in oil have indicated complete recovery when the oil is injected directly into the furnace. In this paper, the problems that cause difficulties in species determination of organometallic compounds have been discussed. H P L C is a very powerful technique for t h e nondestructive separation of organometallic compounds. However,

392

A N A L Y T I C A L CHEMISTRY, VOL. 51, NO.

3, MARCH 1979

if this technique is used, the problems of interference caused by the organic eluent and losses of highly volatile organometallic compounds still remain. It is very difficult to separate the eluent from volatile organometallic compounds. Our approach is to vaporize the sample in a quasi-closed space and t o raise the temperature of the sample vapor as high as possible to highly dissociate the molecules. The interference caused by background absorption is avoided by utilizing the Zeeman AA technique. When more complete dissociation of molecules is necessary, a dc or rf discharge in the furnace might be effective.

ACKNOWLEDGMENT We thank S. Hanamura of the National Bureau of Standards for helpful discussions and suggestions.

LITERATURE CITED (1) "A Review of the Physiological Impact of Mercurials", EPA-66013-73-022 (1974). (2) A. Nakahara, Cbem. Ind., (Japan), 30, 153 (1977). (3) Fed. Regist., 38 (6), 1254 (Jan. 10, 1973). (4) S. Hanamura, Private communications (1978). (5) J. C. Van Loon, B. Radziuk, N. Kahn, J. Lichwa, F. J. Fernandez. and J. D. Kerber, At. Absorp. Newsl., 16, 79 (1977). (6) D. R. Jones and S. E. Manaham, Anal. Cbem., 48, 502 (1976). (7) F. E. Brinckrnan, W. R. Blair, K. L. Jewett, and W. P. Iverson, J. Cbromatogr. Sci., 15, 493 (1977).

(8) H. Koizurni, T. Hadeishi, and R. D. McLaughlin, Anal. Chem., 50, 1700 (1978). T. Hadeishi, Appl. Pbys. Lett., 21, 438 (1972). T. Hadeishi and R. D. McLaughlin, Anal. Cbem., 48. 1009 (1976). H. Koizumi and K. Yasuda, Spectrochim. Acta, Part B , 31, 237 (1976). H. Koizumi, K. Yasuda, and M. Katayama, Anal. Chem.,49, 1106 (1977). J. W. Robinson, Anal. Cbim. Acta, 24, 451 (1961). R. E. Sturgeon and C. L. Chakrabarti. Spectrochim. Acta, Par? 6 , 32, 231 (1977). (15) R. E. Sturgeon and C. L. Chakrabarti, Anal. Chem.. 49, 1100 (1977). (16) B. V. L'vov, Talanta, 23, 109 (1976). (17) H. Massmann, Spectrochim. Acta, Part 8,23, 215 (1968). (18) R. Woodriff and G. Ramlow, Spectrochim. Acta, PartB, 23, 665 (1968). (19) R . Woodriff and R. Stone, Appl. Opt., 7 , 337 (1968). (20) "Handbook of Chemistry and Physics", 53rd ed., R. C. Weast, Ed., The Chemical Rubber Co., Cleveland, Ohio, 1973. (21) R. E. Sturgeon, C.L. Chakrabarti, and C. H. Langford, Anal. Chem., 48, 1792 (1976). (22) D. T. Coker, Anal. Chem., 47, 386 (1975). (23) R. B. Cruz and J. C. Van Loon, Anal. Chim. Acta, 72, 231 (1974). (24) E. J. Czobik and J. P. Matousek, Anal. Chem., 50, 2 (1978). (25) M. D. Amos, P. A. Bennett, K. G. Brodie, P. W. Y. Lung, and J. P. Matousek, Anal. Chem., 43, 211 (1971). (26) C. W. Fuller, At. Absorp. News/., 18, 106 (1977). (27) D. C. Manning and W. Slavin. At. Absorp. Newsl., 17, 43 (1978). (28) F. J. Fernandez and D. C. Manning, At. Absorp. News/., I O , 65 (1971). (29) J. S. Verbeke, Y. Michotte, P. V. Winkel, and D. L. Massart. Anal. Chem., 48, 125 (1976). (30) J. Aggett and A. J. Sprott, Anal. Cbim. Acta, 72, 49 (1974). (9) (10) (1 1) (12) (13) (14)

RECEILTD for review July 21, 1978. Accepted October 10, 1978.

Gas-Liquid Chromatographic Study of the Thermodynamics of Solution in Dipropionitriles R. K. Kuchhal and

K. L.

Mallik"'

Indian Institute of Petroleum, Dehradun-248005, India

Gas-liquid chromatography has been employed to study the thermodynamic solution properties of a number of hydrocarbons and oxygenated compounds in @,@'-oxydipropionitrile, b',P'-thiodipropionitrile, and p,P'-iminodipropionitrile. Comparative data on these three solvents are reported for three temperatures 40, 50, and 60 'C. The data in TDPN for some solutes are compared with previously published data. The activity coefficient data are presented after correction for surface effects (solute adsorption at the gas-liquid interface). The Y ~ , data , ~ are correlated with a linear relation between the partial molar excess free energy and the hydrocarbon chain length of the solute for a homologous series in a given solvent. The results are also discussed in terms of statistical approaches (thermal and athermal contribution) and partial excess molar thermodynamic quantities.

T h e high selectivity for aromatic hydrocarbons over paraffins exhibited by propionitriles like @,@'-oxydipropionitrile (ODPN), 0,B'-iminodipropionitrile ( I D P N ) , and d,@'-thiodipropionitrile (TDPN) was first indicated by Saunders ( 1 ) and subsequently further explored by Medcalf et al. ( 2 ) in their studies on ODPN as a selective solvent for aromatic recovery by solvent refining. Soon after the introduction of the gas chromatographic technique in the mid-fifties, these nitriles, especially ODPN, had gained wide popularity and have been 'Present address: Research &: Development Centre, Indian Oil Corporation Limited, Sector-13, Faridabad-121002, Haryana, India. 0003-2700/79/0351-0392$01 0010

in use as stationary phase in gas-liquid chromatography (GC) as evidenced in the subsequent numerous publications (3-13). In the intervening years, with gradual perfections in the technique (141, GC has provided the chemist with a rapid and convenient means to obtain a large volume of data on solute-solvent interactions, which was useful for studying the thermodynamics of solutions. T h e activity coefficient d a t a determined by the GC technique have been found to be thermodynamically reliable and the data compare quite favorably with those extrapolated from static measurements under well-defined equilibrium conditions. T h e objective of the present work is to provide carefully determined GC equilibrium data on the three nitrile stationary phases over a limited temperature range for diverse types of solutes. Although, as stated earlier, some workers have used these solvents in GC for several purposes, systematic studies on the thermodynamics of solution on these nitriles have not been carried out. Utilizing the solute-solvent interaction information available through the GC technique, i t was felt interesting t o study the changes in thermodynamic parameters, if there are any, with the systematic change in the structure of the solvents with introduction of different groupings such as 0