Anal. Chem. 1989, 6 1 , 1175-1178
1175
TECHNICAL NOTES Application of Signal Enhancement by Easily Ionized Elements in Hydride Generation Direct Current Plasma Atomic Emission Spectrometric Determination of Arsenic, Antimony, Germanium, Tin, and Lead Ian D. Brindle* and Xiao-chun Le Chemistry Department, Brock University, St. Catharines, Ontario, Canada L2S 3A1
INTRODUCTION Signal enhancement by easily ionized elements (EIE) in direct current plasma atomic emission spectroscopy (DCPAES) has been encountered for over 2 decades (1-4). In the presence of easily ionized elements, such as alkali elements and alkaline-earth elements, analytical signals of other elements are usually increased. Although emission line to background ratios are often improved as well, this enhancement effect has been generally regarded as nuisance (2-15) due to its unpredictability. For DCP-AES determinations, variable EIE concentrations in different sample matrices make the accurate determination difficult. Even for fixed E I E concentrations, emission line enhancements differ between wavelengths of the same element and between different elements (1, 16). In order to reduce the interference caused by EIE, a number of possible approaches for elemental analyses have been suggested. These have included matrix matching of samples, blanks, and standards (2,6, 7), standard additions or internal standards ( 8 , 9 ) ,E I E buffers (9-14), and removing EIEs by ion exchange (15). Among these approaches, E I E buffering has been most often applied, where samples are typically spiked with excess amounts of EIEs, such as salts of lithium (IO, I I ) , sodium (9, 12), potassium (9, 13), cesium (14), and lithium-lanthanum mixture (5)to minimize signal differences due to EIEs present in samples with unknown matrices. However, these compensations and matrix matches have not always been successful since enhancement effects due to EIEs are very complicated ( I ) . Through the study of characteristics and possible mechanism, a number of papers have ascribed the signal enhancement effect by EIEs to suppression of populations of ionized analyte by EIE-donated electrons (17-19). Later research, however, showed that intensities of both atomic and ionic emission lines were enhanced and that the addition of E I E appeared to have little influence on electron densities measured in the analytical zone (16, 20, 21). With more detail study, a number of mechanistic models have suggested possible explanations of this complex phenomenon (3,8,22,23). An important paper by Miller et al. (22),proposes a radiative transfer-collisional redistribution of energy model for analyte excitation. Hydride generation has been shown a n efficient sample introduction technique for trace analysis (24-26). Through this process, an analyte is simultaneously concentrated as its hydride and separated from the sample matrix. Therefore, sensitivity and detection limit are improved and spectral interference is reduced. If the hydride is introduced to the dc plasma while a solution of EIE with constant concentration is separately nebulized and transported to the plasma jet, it should be possible to obtain an improved signal-to-background ratio which is beneficial to trace analysis. Hydride generation
* To whom correspondence should be addressed.
coupled with dc plasma is an ideal situation where interference by EIE could become an advantage by increasing the sensitivities and improving the detection limits of hydride-forming elements. With our previous modification of the dc plasma sample tube for hydride introduction (27),the experiment based on this idea is easily performed. Previous studies (26, 28) showed that with conventional aspiration and nebulization of distilled water through the outer tube into the plasma while the hydride was introduced through the inner tube, the dc plasma was more stable and better base line was achieved although the sensitivity remained unchanged. The replacement of distilled water by solutions of EIEs, it was felt, could allow the signal enhancement effect of EIEs to advantage. Thus we report a series of investigations on the determination of antimony, arsenic, germanium, lead, and tin by hydride generation coupled with direct current plasma atomic emission spectrometry.
EXPERIMENTAL SECTION Apparatus. The equipment used included a Spectraspan V dc plasma atomic emission spectrometer (Spectrametrics) with a modified sample tube, a Dataspan data storage system, a Sargent-Welch XKR chart recorder, and a Beckman hydride generator as described previously (27) with the following further modifications. For the determination of tin (26)and germanium (28), only a CaSO, drying tube was used instead of a Porapak Q delay column and a calcium chloride drying column. For the determination of arsenic, antimony, and lead, a water trap was used instead of the drying column and delay column. A Brinckman variable volume Macro-Transferpettor was used for all analyte injections with the volume fixed at either 5.0 or 9.0 mL. A disposable syringe was used for the injection of sodium tetrahydroborate(II1) solution. An entrance slit size of 50 pm (horizontal) and 300 pm (vertical) and an exit slit size of 100 I.cm (horizontal) and 300 pm (vertical) were chosen as recommended (29)for all five elements studied. Photomultiplier voltage and gain settings were used appropriately to provide the best signal to noise ratio. For the determination of arsenic, antimony, germanium, lead, and tin, the wavelengths of 193.696, 259.805, 303.906, 283.306, and 283.999 nm, respectively, and carrier gas flow rates of 970, 970, 740, 930, and 1210 mL min-', respectively, were used. Reagents. Tin (26)and germanium (28)standard solutions and arsenic(II1)and arsenic(V) stock solutions (30)were prepared as described previously. Standard solutions of arsenic were made in either 0.10 M HNO, or 0.10 M HN03-1% (m/V) L-cysteine. Antimony stock solution (1000.0 mg L-I) was prepared by dissolving an appropriate amont of Sb203(Baker Analyzed Reagent) in 2.0 N HCl and diluting with deionized water to 1000.0 mL. Standard solutions of antimony were prepared by serial dilutions of this stock solution with 0.050 M HNO,. Lead stock solution (lOOO.0 mg L-I) was prepared by dissolving an appropriate amount of accurately weighed Pb(N03), (AESAR, 99.999%) in 2.0 N HNO, and diluting to 1OOO.O mL with deionized water. Standard lead solution was made in 0.30% (m/V) potassium dichromate and 0.060 M malic acid. Analytical reagents LiOH, NaC1, KC1, CsC1, MgCl,, CaCl,, SrCl,, and BaC1, (BDH AnalaR), LiCl and Li2S04(Fisher Sci-
0003-2700/89/0361-1175$01.50/00 1989 American Chemical Society
1176
ANALYTICAL CHEMISTRY, VOL. 61, NO. IO, MAY 15, 1989
A
D C Plasma Jet
Figure 1. Schematic representation of the introduction of hydrides and EIEs to the dc plasma: A, hydride introduction; B, modified sample tube; C, nebulizer chamber; D, pneumatic chamber; E, peristaltic pump; S, introduction of distilled water or EIEs; W, to waste.
entific), RbCl (AESAR), and MgS0, (Baker) were prepared by weighing appropriate amounts of the above reagents and dissolving in distilled water to make cation concentrations of 1.0 or 2.0 M. Dilute EIE solutions were prepared by diluting the above solutions with distilled water. Sodium tetrahydroborate(II1) (Anachemia) solutions (4.070, 6.0%, 10.0%)were prepared in 0.10 M NaOH and were filtered before use. For tin determinations, the NaBH, solution was sparged with argon to reduce the tin blank (26). L-Cysteine, L-cystine, and DL-malic acid (Sigma) and K,Cr,O, (Fisher) were analytical grade. Introduction of Hydride and EIEs. As shown in Figure 1, a modified sample tube (B),discussed elsewhere (27), was used to introduce hydride, carried by argon from the hydride generator, to the dc plasma. The inner tube of the modified sample tube was designed for introduction of gaseous hydrides (A), while the outer tube was originally designed for the introduction of analyte solution for conventional DCP-AES. In this work, the latter was used for the introduction of distilled water and EIE solutions. The Spectraspan peristaltic pump (E) was used to pump distilled water into the nebulizer (D and C) and thence around the outer tube of the sample tube. Therefore, a simultaneous but separate introduction of gaseous hydrides and EIE solutions was achieved. Distilled water and EIE solutions were pumped at a rate of 1.9 mL min-'. Normal conditions of nebulizer gas flow (29) were used such that the nebulizer pressure was kept at 26 psi. Hydride Generation Procedures and Parameters. A gas flowing batch system was used for hydride generation. Argon was allowed to flow continuously through the reaction vessel of the hydride generator during the entire working time. The analyte solutions were added, followed by sodium tetrahydroborate(II1) solution except for the antimony determination where NaBH, was added prior to the antimony analyte solution. Because argon was kept flowing through the reaction vessel while the analyte solution and NaBH4 were added, argon stripped the hydrides and then carried them to the plasma. Atomic emission signals were then obtained by the echelle spectrometer. After the determination was completed, the reduced solution was drained from the reaction vessel to waste. The reaction vessel was washed 3 times with deionized water before the next determination was performed. The basic parameters of hydride generation procedures for As, Sb, and P b are summarized below. Methods for the determination of tin (26) and germanium (28)by hydride generation have been discussed elsewhere. Arsenic. Although previous work on arsenic determination (30) was satisfactory, especially in reducing interferences, the method required high acid concentration (5 M HCl) and 30 s of reaction
time. In order to speed the determination, a gas flowing batch system (26) was adopted. Preliminary results showed that as the HN03 concentration increases from 0.01 to 0.20 M, the peak heights of arsenic signals increase. High acid concentration may cause a distortion of plasma due to the large amounts of hydrogen produced. A nitric acid concentration of 0.10 M was chosen, where sufficient arsenic signal was obtained. Further, addition of Lcysteine to give a concentration of 1.0% (m/V) in the arsenic solution was found to speed the reaction and reduce interference. Therefore, 0.10 M HNO, and 1%L-cysteine were used to prepare both arsenic(II1) and arsenic(V) testing solutions. Analyte solution (5.0 mL) and 1.5 mL of 6.0% (m/V) sodium tetrahydroborate(II1) solution were used for each determination. Antimony. A gas flowing batch system was also used for antimony determination. With the study of acid concentration on antimony determination by hydride generation, the highest antimony signal was obtained at 0.050 M HNO,; therefore, 0.050 M nitric acid was used for the preparation of the antimony solutions. It should be noted that even a very little NaBH4 remaining in the reaction vessel left from a previous determination can reduce antimony to its hydride. In order to avoid the uncertainty caused by unreacted NaBH,, a reversed order for injection of sample and sodium tetrahydroborate(II1) solutions was used, i.e., injection of sodium tetrahydroborate(II1) (1.0 mL of 6.0% (m/V)) prior to the sample solution (5.0 mL) to ensure that an excess of NaBH, was available. The antimony signal was observed immediately after the antimony solution was injected. Lead. The conditions for lead determination by hydride generation have been found to be very critical (31). Three reaction media, (NH4)2S208-HN03,H202-HCl, and K2Cr207-malic acid, were tried for lead determination. Although the (NH,),S,O,-HNO3 system gave the highest lead sensitivity, it was more prone to interference from foreign ions, especially transition-metal ions (31). It is found that with K2Cr207-malic acid as the hydride formation medium for lead, the least interference was observed. Preliminary experiments showed that 0.3 % (m/V) potassium dichromate-0.060 M malic acid and 10.0% (m/V) sodium tetrahydroborate(II1) were the optimum concentrations for lead determination. The volumes of analyte solution and sodium tetrahydroborate(II1) solution used for lead determination were 5.0 and 1.5 mL, respectively.
RESULTS AND DISCUSSION Enhancement of Analyte Signal and Changes of Background. Signal enhancement by EIEs is an important feature in DCP-AES. It depends on different EIEs (3, I O ) , concentration of EIE compounds ( 5 , 11, 12), and different elements determined (8,10,22). With the concern for these factors, this effect was studied by introducing each of the alkali and alkaline-earth element-containing solutions at various concentrations in the determination of As, Sb, Ge, Sn, and P b by hydride generation DCP-AES. Distilled water, alkali, and alkaline-earth element-containing solutions, with concentration ranging from 0.10 to 2.0 M, were introduced to the d c plasma through a conventional peristaltic pump and a pneumatic nebulizer while the hydride, carried by argon, was introduced through the inner tube t o the dc plasma. Atomic emission spectroscopic signals were then obtained in duplicate with the spectrometer and recorded with a chart recorder. On comparison of the peak heights of the emission signals obtained in the presence of EIE solutions to those in the presence of distilled water, signal enhancement data and background results were calculated and are shown in Table I. The background increase (+) and reduction (-) values indicate the background changes in the presence of EIEs compared to the background in the presence of distilled water, where the latter was between 300 and 600, depending on the elements determined and the photomultiplier tube and gain settings used, while the full scale of recording was 10000 arbitrary units. T h e concentrations of As, Sb, Ge, Sn, and P b solution used in performing this experiment were 20.0, 20.0, 2.0, 2.0, and 20.0 ng mL-', respectively. As shown in Table I, in all cases the As, Sb, Ge, Sn, and
ANALYTICAL CHEMISTRY, VOL. 61, NO. 10, MAY 15, 1989
1177
Table I. Effect of EIEs on Signal and Background in the Determination of As, Sb, Ge, Sn, and Pb"
EIE compound
LiCl LiOH
Li2SO,
NaCl
EIE concn, M 1.0
2.0 0.10 0.50 1.0 2.0 0.10 0.50 1.0 2.0 0.10 0.50 1.0
KC1
RbCl CSCl
2.0 0.10 0.50 1.0 2.0 0.50 1.0 2.0 0.50 1.0 2.0 1.0 2.0 1.0
2.0 1.0 2.0 1.0 2.0 0.50 1.0
% signal enhancement
As
Sb
Ge
24 24
46
-
-
-
17
5 13 15 22 5 15 20 23 23 44 44 56 40 57 78 103
28 38 57 61 24 41 45 55 39 52 65 83 28 52 66 76
56 103 64
46 68
49 57
-100
-
5 12 21 29 12
19 21 31 5 21 26 19
-
29 39 42 66 17 24 46 61 12
36 36 31 41 19 26 36 29 36 38 36 38
63 61 3 51 61 10 39 76 15 66 66 68 76 56 63 42
19 19
37 39
21
-
-
-
-
-
43 54 49 49 29 43
Sn
-
-55
-
-25 -55 0
-26
Pb
60 70 26 56 60 70 37 54 61 73 40 50 60 65 30 50 70 88 -
90 90 85 90 115 -59 -
-69 -
73 73 95 73 77 50
background increase (+) or reduction (-) based on a background signal of 300-600, '70 As Sb Ge Sn Pb -19 -20 -
-
n.d. -10
-14 -17 n.d. -12
-19 -18 n.d. -19 -26 -26 -18 -30 -30 -33 -34 -33 n.d. -6 n.d. n.d. -10 -11 n.d. -
-7 -9
-27 -25 +13
-
-
-
-
n.d. +6 +16
n.d. n.d. -6
-
+7 +14 +16 +18 n.d.
n.d.
+6
+10 n.d.
+11 +8
n.d.
+16 +16 +19
-11
+22 +9 n.d. -21
+36 n.d. -21
-41 +21 -20 -31 -37 -34 -42 +78 +59 +70 +73 +60 +77 +74 -
+110 +110
-
+8
+8
n.d. n.d. n.d. n.d.
n.d. n.d.
-
n.d. n.d. n.d. n.d. n.d. n.d. n.d.
n.d. n.d. n.d.
+16 +16 +21
n.d. -6
+6 +10 n.d.
+31 +63
+7 -10 -10 -14 -18 +1500
-
+150
+260
-
+110 +360 +130 +450 +320 +580
-
-
+1400
-
+270
-
+180 +300 +150 +290
n.d. -
-
+1500
-
+140 +250 +130 +220 +200 +340
Key: n.d., not detectable; -, not determined, P b , analytical signals were significantly enhanced with any of the alkali-element-containing compounds. Comparing the results obtained with aspiration of LiC1, LiOH with Li,S04, and MgClz with MgS04, we can see that the signal enhancement effect is essentially independent of anion b u t depends on the cation. With the increase of alkali-metal concentration from 0.10 to 2.0 M, the signal enhancement factor increases. Although a higher concentration of alkali elements may give a higher signal enhancement factor, a concentration of E I E of 1.0 M is recommended since a higher salt concentration may cause solid salt to appear on the top of the sample tube. I t is worthy of note t h a t although the background level was changed with the introduction of alkali elements, no significant change of noise was observed. With aspiration of alkaline-earch-element-containing solution to the dc plasma as E I E source, the signal and background results are more complicated and depend on the different analytes. As we can see in Table I, in arsenic determinations, signals were enhanced and the background was decreased with all alkaline-earch elements, whereas the opposite phenomenon was observed in the case of tin determinations. In the determination of antimony and germanium, both signal and background were increased with increasing alkaline earth element concentration. For the determination of lead, Ca, Sr, and Ba enhanced the lead signal, whereas Mg suppressed the lead signal. As a result of enhanced intensities of emission lines ( 1 , 3 ) , both peak height and peak area of emission signals were increased. In the cases of lead and germanium, for example, the signals were enhanced by 70% and 78%, respectively, when 1.0 M KC1 was introduced compared to the normal
situation where distilled water was introduced. As one might predict, the enhancement effect of the E I E does not depend on the hydride generation procedure. As a trial, arsenic(II1) and arsenic(V) were used as standards in solution with and without L-cysteine. T h e effect of EIEs on arsenic signals was then studied with the above reaction solutions. The results showed the same signal enhancement was observed by using either As(II1) or As(V) as reaction solution. Also, similar enhancement effects were obtained in the presence and absence of L-cysteine, although L-cysteine speeded the hydride reaction and reduced the interference in the hydride generation process. Stability of the Enhanced Signals. In conventional DCP-AES, the uncertainty and nonreproducibility caused by EIEs are severe problems due to variable unknown E I E concentrations in different samples and to the difficulty in matching the matrix of sample and standard solutions. However, this problem is overcome in the hydride generation dc plasma system since the hydrides were separated from the sample matrix. Thus alkali or alkaline-earth elements were not introduced to the plasma from the hydride generator along with hydride even though they were present in the sample. Since separate introduction of E I E solutions through the peristaltic pump and pneumatic nebulizer can be easily controlled, the concentration of EIE in the plasma is kept constant. As shown in Figure 2, very reproducible germanium signals were obtained with and without EIE. Eleven replicate determinations of 0.50 ng mL-' germanium gave a relative standard deviation (RSD) of 4.7% fer introduction of distilled water and 4.1% for introduction of 1.0 M KC1. Similar results were obtained for lead determination, where 11 replicate
1170
Anal. Chem. 1989, 61, 1178-1181 B
C
other plasma sources since the dc plasma can tolerate high levels of salts without adversely affecting the plasma. Detection limits for As, Ge, Sb, Sn, and P b determinations were 230, 12, 200, 20, and 140 pg mL-'. Further investigation of the optimum conditions for the determination of As and Sb should lead to improvements in the detection limits for these two elements.
LITERATURE CITED
Comparison of signals of blank and germanium in the presence of 1.0 M KCI and distilled water: A, blank, with introduction of 1.0 M KCI; B, 0.50 ng mL-' Ge, with introduction of 1.0 M KCI; C, 0.50 ng mL-' Ge, with introduction of distilled water. Flgure 2.
determinations of 2.0 ng mL-' lead gave RSD of 6.0% in the absence of EIE and 4.6% in the presence of 1.0 M CsC1. We can also see from Figure 2 that there was no signal from the germanium blank even in the presence of 1.0 M KCl. The introduction of EIE reagents, with analytical grade purity, to the dc plasma did not increase the blank value because the concentrations of the analytes in these reagents are negligible compared to their detection limits by conventional DCP-AES method. Calibration. With EIEs, reproducible and enhanced signals were achieved, which is promising in improving sensitivity for trace analysis. In order to estimate the analytical applicability of signal enhancement by EIE, selective calibration and linear regression were performed. In the cases of lead, for example, correlation coefficients of two curves were 0.997 for water aspiration and 0.998 for 1.0 M CsCl aspiration over the range of 0-80.0 ng mL-'. The slopes of the line were 1.14 and 2.17 for the introduction of distilled water and 1.0 M CsC1, respectively. Detection Limits. Detection limits, defined as 3 times the noise, for the determination of arsenic, antimony, germanium, and lead, were 230,200,12, and 140 pg mL-', respectively, with the introduction of 1.0 M EIE solution (KC1 or CsCl), whereas they were 360, 360, 20, and 250 pg mL-', respectively, in the absence of EIE. The detection limit for the determination of tin was limited by the reagent blank to a level of 20 pg mL-' as previously reported, although the introduction of EIE further enhanced the signal.
CONCLUSION Emission enhancement by EIEs has shown advantages in improving sensitivities for the determination of As, Sb, Ge, Sn, and P b by hydride generation. Alkali elements are preferred to alkaline-earth elements since the latter gave much higher background. The application of EIE signal enhancement seems to be more practical with dc plasma than with
Miller, M. H.; Eastwood, D.; Hendrick, M. S.Spectrochlm.Acta, Part8 1984, 398, 13-56. Szivek, J.; Jones, C.; Paulson, E. J.; Valberg. L. S. Appl. Spctrosc. 1968, 22. 195-197. Nygaard, D. D.; Gilbert, T. R. Appl. Spectrosc. 1981. 35,52-56. Johnson, G. W.; Taylor, H. E.; Skogerboe, R. K. Anal. Chem. 1979, 51. 2403-2405. Fox, R. L. Appl. Spectrosc. 1984, 38, 644-647. Cantillo, A. Y.; Sinex. S. A.; Helz, G. R. Anal. Chem. 1984, 5 6 , 33-37. Biggs, W. R.; Fetzer, J. C.; Brown, R. J. Anal. Chem. 1987, 59, 2798-2802. Johnson, G. W.; Taylor, H. E.; Skogerboe, R. K. Appl. Spectrosc. 1980. 3.4 . 19-24. ..., Nygaard, D. D.-Anal. Chem. 1979, 5 1 , 881-884. Skogerboe, R. K.; Urasa, I. T. Appl. Spectrosc. 1978, 32, 527-532. Frank, A.; Petersson, L. R. Spectrochim. Acta, Pari 8 1983, 388, 207-220. Urasa, I . T. Anal. Chem. 1984, 56, 904-908. Goiightiy, D. W.; Harris, J. T. Appl. Spectrosc. 1975, 29, 233-240. Bankston, D. C.; Humphris, S. E.; Thompson, G. Anal. Chem. 1979, 51. 1218-1225. Lajunen. L. H. J.; Kurikka, A.; Ojaniemi, E. At. Spectrosc. 1987, 8 , 142- 144. Eastwood, D.; Hendrick, M. S.; Miller, M. H. Spectrochim.Acta, Pari 8 1982. 378. 293-302. Johnson, G. W.; Taylor, H. E.; Skogerboe, R. K. Spectrochim. Acta, Part E 1979, 3 4 8 , 197-212. Williams, R. R.; Coleman. G. N. Appl. Spectrosc. 1981, 35,312-317. Felkel, H. L., Jr.; Pardue, H. L. Anal. Chem. 1978, 50,602-610. Blades, M. W.; Lee, N. Spectrochlm. Acta, Pari B 1984, 398, 879-890 - . Zander, A. T.; Miller, M. H. Spectrochlm. Acta, Part 8 1985, 4 0 8 , 1023- 1037. Miller, M. H.; Keating, E.; Eastwood, D.; Hendrick, M. S. Spectrochim. Acta. Part E 1985. 408. 593-616. Hendrick, M. S.; Seltzer. M. D.; Michei, R. G. Spectrochim, Acta, Part 8 1986, 4 1 8 , 335-348. Robbins, W. B.; Caruso, J. A. Anal. Chem. 1979, 5 1 , 889A-899A. Sparkes, S.;Ebdon, L. ICP I n f . News/. 1986, 12, 1-6. Brindle, I. D.; Le, X-c. Ana/ysf (London) 1988, 113, 1377-1381. Boampong, C.; Brindle, I. D.; Ceccarelli Ponzoni, C. M. J . Anal. At. Spectrom. 1987, 2 , 197-200. Brindle, I. D.; Le, X-C.; Li, X-f. Presented at the 4th Biennial National Atomic Spectroscopy Symposium, York, UK, June 29-July 1, 1988. J. Anal. At. Spectrom. 1989, 4 , 227-232. Spectraspan V Emission Spectrometer, Operator's Manual: Spectrametrics: Andover, MA, 1983. Boampong, C.; Brindle, I . D.; Le, X-c.; Pidwerbesky. L.; Ceccarelli Ponzoni, C. M. Anal. Chem. 1988, 60, 1185-1188. Jin, K.; Taga. M. Anal. Chlm. Acta 1982, 143, 229-236.
RECEIVED for review October 24,1988. Accepted January 31, 1989. The authors gratefully acknowledge receipt of a grant from the Ontario Government BILD program for the purchase of the Spectraspan V dc plasma atomic emission spectrometer. The authors also thank the Air Resources Branch of the Ontario Ministry of the Environment for funding this research (project 360 G ) .
Back-Pressure Regulated Restrictor for Flow Control in Capillary Supercritical Fluid Chromatography Douglas E. Raynie, Karin E. Markides, Milton L. Lee, and Steven R. Goates* Department of Chemistry, Brigham Young University, Prouo, Utah 84602
INTRODUCTION In chromatography, control of mobile phase flow rates near the optimal linear velocity preserves both efficiency and resolution. With gaseous mobile phases, especially in con0003-2700/89/0361-1178$01.50/0
junction with capillary columns, flow rate changes during temperature programming are commonly experienced but are not severe. In liquid chromatography, pumps are operated synchronously to deliver a controlled flow rate. Pressure 1969 American Chemical Society