Anal. Chem. 1995, 67, 2059-2063
`
`Fused Quartz Substrates for Microchip
`Electrophoresis
`Stephen C. Jacobson,* Alvin W. Moore, and J. Michael Ramsey
`Chemical and Analytical Sciences Division, Oak Ridge National Laboratory, P.O. Box 2008
`Oak Ridge, Tennessee 37831-6142
`
`A fused quartz microchip is fabricated to perform capillary
`electrophoresis of metal ions complexed with 8-hydroxy-
`quinoline- 5-sulfonic acid (HQS). The channel manifold
`on the quartz substrate is fabricated using standard
`photolithographic, etching, and deposition techniques. By
`incorporating a direct bonding technique during the
`fabrication of the microchip, the substrate and cover plate
`can be fused together below the melting temperature for
`fused quartz. To enhance the resolution for the separa-
`tion, the electroosmotic flow is minimized by covalently
`bonding polyacrylamide to the channel walls. A separa-
`tion length of 16.5 mm and separation field strength of
`870 V/cm enable separations to be performed in <15 s.
`By increasing the concentration of HQS from 5 mM to 20
`mM, the separation efficiency improves by ~3 times. The
`low background signal from the fused quartz substrate
`results in mass detection limits of 85,61, and 134 amol
`and concentration detection limits of46,57, and 30 ppb
`for Zn, Cd, and Al, respectively.
`
`Miniaturized chemical analysis instruments fabricated using
`micromachining techniques are receiving increasing attention. The
`more successful demonstrations for miniaturized instruments
`incorporate liquid phase separation techniques. Devices have
`been demonstrated for capillary electrophoresis,1-6 free-flow
`electrophoresis,1 234567 open channel electrochromatography,8 910*and capil-
`lary electrophoresis with pre- and postcolumn derivatization.910
`All of the previously demonstrated devices utilize glass substrates.
`By moving to quartz substrates, the superior optical properties
`of quartz especially in the UV region can be exploited. For both
`capillary electrophoresis and liquid chromatography, the most
`useful wavelength range for the purpose of optical detection is
`200-400 nm, and the ability to fabricate fused quartz microchips
`permits investigation of similar detection scenarios.
`
`(1) Harrison, D. J.; Manz, A; Fan, Z.; Lüdi, H.; Widmer,  , M. Anal. Chem.
`1992, 64, 1926.
`(2) Manz, A; Harrison, J.; Verpoorte, E. M. J.; Fettinger, J. C.; Paulus, A; Liidi,
`H.; Widmer, H. M.J. Chromatogr. 1992, 593, 253.
`(3) Seiler, K; Harrison, D. J.; Manz, A Anal. Chem. 1993, 65, 1481.
`(4) Harrison, D. J.; Fluri, K; Seiler, K; Fan, Z.; Effenhauser, C. S.; Manz, A
`Science 1993, 261, 895.
`(5) Jacobson, S. C.; Hergenrtider, R.; Koutny, L. B.; Warmack, R J.; Ramsey, J.
`M. Anal. Chem. 1994, 66,1107.
`(6) Jacobson, S. C.; Hergenrtider, R; Koutny, L. B.; Ramsey, J. M. Anal. Chem.
`1994, 66, 1114.
`(7) Raymond, D. E.; Manz, A; Widmer,  . M. Anal. Chem. 1994, 66, 2858.
`(8) Jacobson, S. C.; Hergenrtider, R; Koutny, L. B.; Ramsey, J. M. Anal. Chem.
`1994, 66, 2369.
`(9) Jacobson, S. C.; Hergenrtider, R; Moore, A W., Jr.; Ramsey, J. M. Anal.
`Chem. 1994, 66, 4127.
`(10) Jacobson, S. C.; Koutny, L B.; Hergenrtider, R; Moore, A W., Jr.; Ramsey,
`J. M. Anal. Chem. 1994, 66, 3472.
`
`0003-2700/95/0367-2059S9.00/0 &copy; 1995 American Chemical Society
`
`injection
`cross
`
`Figure 1. Schematic of the fused quartz microchip. The reservoirs
`are affixed on the microchip via epoxy.
`
`To test the performance of fused quartz substrates, metal ions
`complexed with S-hydroxyquinoline-5-sulfonic acid (HQS) are
`separated by electrophoresis and detected with UV laser-induced
`fluorescence. HQS has been widely used as a ligand for optical
`determinations of metal ions following initial work by Fiegl and
`Heisig.11 The optical properties and the solubility of HQS in
`aqueous media have recently been used for detection of metal
`ions separated by ion chromatography12 and capillary electro-
`phoresis.13 Because uncomplexed HQS does not fluoresce, excess
`ligand is added to the buffer to maintain the complexation
`equilibria during the separation without contributing a large
`background signal. This benefits both the efficiency of the
`separation and the detectability of the sample.
`In this paper, the fabrication of fused quartz microchips is
`described. Direct bonding of the cover plate to the substrate
`enables covalent bonding to be performed below the melting
`temperature of fused quartz. A simple cross channel manifold
`(Figure 1) is constructed to perform electrophoretic analysis of
`metal ions complexed with HQS. The electroosmotic flow of the
`microchip is minimized by covalently bonding polyacrylamide to
`the surface of the channels. This enables a higher resolution to
`be achieved in a shorter separation time and length.
`Improved
`efficiency with increasing HQS concentration is also discussed.
`
`EXPERIMENTAL SECTION
`The microchips are fabricated using standard photolitho-
`graphic, wet chemical etching and bonding techniques.14 First,
`
`(11) Fiegl, F.; Heisig, G. B. Anal. Chim. Acta 1949, 3, 561.
`(12) Soroka, K.; Vithanage, R S.; Phillips, D. A; Walker, B.; Dasgupta, P. K.
`Anal. Chem. 1987, 59, 629.
`(13) Swaile, D. F.; Sepaniak, M. J. Anal. Chem. 1991, 63, 179.
`(14) For example: Ko, W. H.; Suminto, J. T. In Sensors; Gopel, W., Hasse, J.,
`Zemmel, J. N., Eds.; VCH: Weinhein, Germany, 1989; Vol. 1, pp 107-168.
`Analytical Chemistry, Vol. 67, No. 13, July 1, 1995 2059
`
`Downloaded via UNIV OF VIRGINIA on November 6, 2018 at 17:55:00 (UTC).
`
`See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
`
`Agilent Exhibit 1282
`Page 1 of 5
`
`

`

`time [s]
`Figure 3. Electropherogram of Zn, Cd, and Al complexed with
`(a, top) Esep = 870 V/cm and Z.sep = 16.5 mm. The injected
`HQS:
`concentrations are 8.5 (130), 6.7 (60), and 4.3 ppm (160 µ ) for Zn,
`Cd, and Al, respectively (b, bottom) Esep = 720 V/cm and Lsef> = 16.5
`mm. The injected concentrations are 260 (4.0), 210 (1.9), and 130
`ppb (4.9 µ ) for Zn, Cd, and Al, respectively.
`
`The microchip is operated in either a sample loading or
`separation mode. The floating sample loading, described previ-
`ously,4 is used to transport the sample to the injection cross,
`enabling a larger volume sample plug to be injected onto the
`separation column than ivith the pinched sample loading. The
`average sample plug volume for the floating sample loading is
`estimated to be 120 pL compared to 65 pL for the pinched sample
`loading. These volumes are determined by measuring injected
`plug widths 0.1 mm downstream from the injection cross. For
`comparison, the volume of the injection cross is 43 pL. With the
`floating sample loading, the injected plug has no electrophoretic
`bias, but the volume of sample is a function of the sample loading
`time. Because the sample loading time is inversely proportional
`to the field strength used, for high sample loading field strengths
`a shorter sample loading time is used than for low injection field
`strengths. For example, for a sample loading field strength of
`630 V/cm (Figure 3a), the sample loading time is 12 s, and for a
`sample loading field strength of 520 V/cm (Figure 3b), the sample
`loading time is 14.5 s. Both the pinched and floating sample
`loadings can be used with and without suppression of
`the
`electroosmotic flow. To implement the floating sample loading,
`
`20
`
`40
`
`7.6 µ  
`
`I 0
`
`20
`
`o,
`0)
`13
`
`-20
`
`.......
`-40
`
`-20
`
`width [µ  ]
`Figure 2. Profile of separation channel machined on a fused quartz
`substrate.
`
`the fused quartz substrates (50 mm x 25 mm x
`1 mm; Esco
`Products, Inc. R120110) are sputtered with metal films of chro-
`mium (30 nm) and gold (400 nm). A positive photoresist (Shipley
`1811) is then spin-coated on top of the Au film, and the column
`design (Figure 1) is transferred to the substrate using an e-beam
`written photomask (Institute of Advanced Manufacturing Sciences,
`Inc.). Following exposure, the metal films are etched using KI/
`I2 for Au and KsFetCFOe/NaOH for Cr. Finally, the channels are
`etched into the substrate in a dilute, stirred HF/NH4F bath at 50
`°C. Due to the isotropic etch of amorphous materials, the channel
`profile is trapezoidal. Figure 2 shows the cross section of the
`quartz channel measured by a profilometer (Alpha Step 200,
`Tencor Instruments) prior to bonding the cover plate. The
`channel depth is 7.6 µ  , and the channel width at half-depth is
`75 µ  . To form the closed network of channels, a circular cover
`plate (25 mm diameter; Esco Products, Inc. R525000) is bonded
`to the substrate over
`the etched channels. The substrate and
`cover plate are first hydrolyzed in NH4OH/H2O2 at 50 °C, rinsed
`in H20, joined, and then annealed at 1100 °C for 5 h. Cylindrical
`glass reservoirs are affixed on the substrate using epoxy. The
`channel walls are then coated with polyacrylamide to minimize
`electroosmotic flow.15 First, the channels of the microchip are
`filled with a solution of [y-(methacryloxy)propyl]trimethoxysilane
`0.4% (v/v) in water (pH 3.5, adjusted with acetic acid) for 30 min.
`The channels are then flushed with water and filled with a 3%
`(w/v) solution of acrylamide in water with 0.1% (v/v) NJfJffl'·
`tetramethylethylenediamine and 0.1% (w/v) potassium persulfate
`for 30 min. The excess acrylamide solution is removed, and the
`channels are rinsed sequentially with water and separation buffer.
`Column performance and separations are monitored on-
`microchip using a single-point detection scheme via laser-induced
`fluorescence (LIF). An argon ion laser (351.1-363.8 nm, 10 mW;
`Coherent Innova 90) is used for excitation and focused to a spot
`length). The
`on the microchip using a lens (100 mm focal
`fluorescence signal is collected using a 21 x microscope objective,
`followed by spatial filtering (0.6 mm diameter pinhole) and spectral
`filtering (425 nm cutoff; Coming 3-73), and measured using a
`photomultiplier tube (PMT; Oriel 77340). The data acquisition/
`voltage switching apparatus is computer controlled using pro-
`grams written in-house in Labview 3.0 (National Instruments).
`Platinum electrodes provide electrical contact from the power
`supply (Spellman CZE1000R) to the solutions in the reservoirs.
`The compounds used for the experiments are zinc sulfate,
`cadmium nitrate, and aluminum nitrate. The buffer is sodium
`phosphate (60 mM, pH 6.9) with 8-hydroxyquinoline-5-sulfonic
`acid (20 mM for all experiments except that in Figure 5; Sigma
`Chemical Co.). At least 50 mM sodium phosphate buffer is
`needed to dissolve up to 20 mM HQS.
`(15) Hjerten, S. /. Chromatogr. 1985, 347, 191.
`
`2060 Analytical Chemistry, Vol. 67, No. 13, July 1, 1995
`
`Agilent Exhibit 1282
`Page 2 of 5
`
`

`

`Table 1. Data for Figure 3a
`electrophoretic
`mobility (104 cm2 V-1 s-1)
`1.65
`
`metal ion
`
`Zn
`
`Cd
`
`Al
`
`1.49
`
`1.28
`
`plate no.
`8200
`
`6500
`
`3200
`
`resolution
`
`2.26
`
`2.86
`
`a voltage is applied to the sample reservoir with the sample waste
`reservoir grounded and the buffer and waste reservoirs floating.
`After the slowest moving ion, e.g., Al, passes through the injection
`cross, the applied potentials are reconfigured for the separation
`In the separation mode, the voltage is applied to the buffer
`mode.
`reservoir with the waste reservoir grounded and with sample and
`sample waste reservoirs at 60% of the buffer reservoir potential.
`Having a voltage applied to the sample and sample waste
`reservoirs at a fraction of the buffer reservoir voltage prevents
`excess sample from bleeding into the separation column.
`
`RESULTS AND DISCUSSION
`Figure 3 shows the separation of three metal ions complexed
`with 8-hydroxyquinoline-5-sulfonic acid. All three complexes have
`a net negative charge. With the electroosmotic flow minimized
`by the covalent bonding of polyacrylamide to the channel walls,
`negative potentials relative to ground are used to manipulate the
`In panels a and
`complexes during sample loading and separation.
`b of Figure 3, the separation channel field strength is 870 and
`720 V/cm, respectively, and the separation length is 16.5 mm.
`The volume of the injection plug is 120 pL, which corresponds to
`16, 7, and 19 fmol injected for Zn, Cd, and Al, respectively, for
`In Figure 3b, 0.48, 0.23, and 0.59 fmol of Zn, Cd, and
`Figure 3a.
`Al, respectively, are injected onto the separation column. The
`average reproducibility of the amounts injected is 1.6% relative
`standard deviation (rsd) as measured by peak areas (six replicate
`analyses). The stability of the laser used to excite the complexes
`is ~1% rsd.
`The primary purpose of eliminating the electroosmotic flow is
`to enhance the resolution of the separation. For electrophoretic
`the resolution between two components can be
`separations,
`written16
`
` JÑl Ph
`t1 B
`4 v“av + ,WEO
`
`(1)
`
`where N is the average number of theoretical plates, µ\ and µß
`are the electrophoretic mobilities for components A and B,
`respectively, µ\\ is the average electrophoretic mobility for A and
`is the electroosmotic mobility. Electroosmosis does
`B, and µ  
`not contribute to the separation. To maximize the resolution,
`electroosmosis must be eliminated or precisely controlled, e.g.,
`application of an external field.17 For simplicity, electroosmosis
`is minimized by covalently bonding polyacrylamide to the channel
`surface.
`Table 1 lists the efficiencies and resolution (eq 1) generated
`for each complex in Figure 3a. The resolution obtained between
`
`(16) Jorgenson, J. W.; Lukács, K. D. Anal. Ckem. 1981, 53, 1298.
`(17) Lee, C. S.; Banchard, W. C.; Wu, C. T. Anal. Chem. 1990, 62, 1550.
`
`separation field strength [V/cm]
`Figure 4. Variation of linear velocity for Zn (O), Cd ( ), and Al ( )
`and separation column current ( ) with separation field strength. Lines
`represent linear fits for separation field strengths from 140 to 580
`V/cm and extrapolated to 870 V/cm.
`
`Zn/Cd and Cd/Al is 2.25 and 2.86, respectively. To obtain a
`comparable resolution, e.g., 2.25, with electroosmotic flow would
`require a longer separation column and analysis time. For
`example, with an electroosmotic mobility of 5.7  
`10™4 cm2 V"1
`s-1 (measured on a microchip without the electroosmotic flow
`minimized), a field strength of 870 V/cm and a resolution of 2.25,
`a separation length of 11.5 cm is required to resolve these three
`metal ion complexes, and the total analysis time increases to 33
`s. Also, to produce the equivalent separation field strength, a 9.9
`kV power supply is required compared to 1.4 kV for this
`separation. The smaller power supply is more amenable to
`miniaturization.
`The analysis time can be reduced by increasing the separation
`field strength and, consequently, the linear velocity of the sample16
`u = µ.^
`(2)
`
`where µ is the electrophoretic mobility and  ^  is the separation
`field strength. The electrophoretic mobility depends on the
`If the heat generated in the separation
`viscosity of the buffer.
`column is not dissipated efficiently during a separation,
`the
`viscosity of the buffer decreases, leading to higher than predicted
`linear velocities for the sample zones.
`In Figure 4 the linear
`velocity of the sample zones are plotted versus
`the electric field
`strength. For separation field strengths of > 720 V/cm, a deviation
`from linear behavior is observed. Also, plotted in Figure 4 is the
`current measured in the separation column. As expected, the
`current in the separation column exhibits the same nonlinear
`behavior for field strengths of >720 V/cm. For separation field
`strengths of 720 and 870 V/cm, the microchip required ~2 min
`for the linear velocities to stabilize. The reproducibility of the
`linear velocities is 0.1% rsd (six replicate analyses). To reduce
`this heating effect at comparable separation field strengths,
`channels with smaller cross-sectional areas or lower concentration
`(lower conductivity) buffers are two possibilities. Optimum HQS
`concentration for detection and separation (discussed below) is
`least 50 mM
`determined to be 20 mM, which requires at
`phosphate buffer for dissolution. Thus, 80 mM for a total buffer
`the operating limit for these experiments.
`concentration is near
`
`Analytical Chemistry, Vol. 67, No. 13, July 1, 1995
`
`2061
`
`Agilent Exhibit 1282
`Page 3 of 5
`
`

`

`Table 2. Detection Limits
`concentration
`ppb
`46
`57
`30
`
`metal ion
`Zn
`Cd
`Al
`
`µ 
`0.71
`0.51
`1.12
`
`120 pL injection volume.
`
`mass0
`(amol)
`
`85
`61
`134
`
`,
`
`2Pm
`/t£sep
`
`dominant form for Zn and Cd is a 2:1 ligand-metal complex, and
`for Al a 3:1 complex. The Al complex with a 3- charge might
`If two or more species exist
`undergo an ion-pairing phenomena.
`in the sample band for Al, additional band dispersion results
`because the species will migrate at different velocities, yet cannot
`be resolved under these conditions. For HQS concentrations of
`>20 mM, significant
`improvements in the efficiency are not
`observed. The relationships between HQS concentration, the
`complex formation constant, and band broadening associated with
`HQS-metal complexes in capillary electrophoresis have been
`investigated.12 Additional band broadening was previously at-
`tributed to nonequilibrium conditions at injection, various HQS-
`metal species in the sample band, and competing equilibria with
`supporting electrolyte.12 For these experiments, nonequilibrium
`conditions at
`injection do not occur because the sample and
`separation buffers are the same. With the magnitude of the
`formation constants listed above and the large excess of HQS,
`multiple metal-HQS species seems improbable.
`For electrophoresis, the contributions to the total plate height
`from the injection plug length, detector observation length,20 axial
`diffusion,21 and Joule heating22 are
`, 4t
`4j
`16¿sep+16¿sep
`
`+
`
`7 X lQ"9<ep^-2^
`Dy
`
`,d)
`
`Figure 5. Variation of plate height for Zn (O), Cd ( ), and Al (O)
`with concentration of HQS. Error bars are ± 
`for three analyses.
`
`In addition to resolving power and analysis time, detection must
`also be considered in an analytical procedure. Detection limits
`are measured using serial dilutions of a stock solution of the metal
`ions. The measurements are performed by separating the mixture
`of the metal ions, using the experimental conditions in Figure
`3b. The data are then extrapolated to a signal-to-noise ratio (S/
`N) of 2, to give the concentration and mass detection limits listed
`in Table 2. Figure 3b shows the electropherogram at the lowest
`concentration injected for the mixture during the detection limit
`study. The peaks correspond to an average S/N for 3 runs of
`16.4,10.7, and 8.6 for Zn, Cd, and Al, respectively, which are within
`1 order of magnitude of the estimated detection limits (Table 2).
`A pH 6.9 separation buffer is near
`the optimum conditions for
`fluorescence detection of the complexes with Zn, Cd, and Al.11
`Detection limits are improved by using a fused quartz substrate
`as opposed to a glass substrate. The observed background signal
`from the experimental apparatus with the fused quartz substrate
`inserted is 30% of the background signal observed for glass. The
`primary contributions to the background signal include stray room
`light, scattered laser light, and inelastic scattering from the
`microscope objective and microchip substrate. Further reductions
`in background signals are expected by improving the quality of
`the optical components.
`A substantial decrease in the plate height is observed with
`increasing concentrations of HQS in the separation buffer. Figure
`5 shows the variation of the plate height with concentration of
`the HQS. The Al complex exhibits the poorest efficiency. The
`formation constants for Zn, Cd, and Al are   16 2, 1014·2,18 and 109,19
`respectively. Because of
`the large formation constants,
`the
`
`(18) Lange’s Handbook of Chemistry, 14th ed.; Dean, J. A, Ed.; McGraw-Hill Book
`Co.: New York, 1992.
`
`2062 Analytical Chemistry, Vol. 67, No. 13, July 1, 1995
`
`where l¡n¡ is the injection plug length, Idet is the detector observation
`length, Lgep is the separation length, Dm is the diffusion coefficient,
`dc is the channel depth,23   is the molar conductivity of the buffer,
`is the thermal conductivity of
`C is the buffer concentration, and  
`the buffer. The contributions from the injection plug length and
`detector observation length assume Gaussian distributions. The
`lengths of the injection plug and the detector observation are
`If these time-independent contribu-
`constant for all experiments.
`tions predominate in their contribution to the plate height, then
`the total plate height decreases as the separation length increases.
`The contribution from axial diffusion to the plate height is reduced
`by increasing the separation field strength and, consequently,
`reducing the analysis time. The contribution from Joule heating
`can be minimized by controlling the channel dimensions, separa-
`tion field strength, and buffer concentration.
`In Figure 6, the variation of the total plate height with the
`electric field strength is plotted. For these experiments,
`the
`injection plug length is 330 µ   and the detector observation length
`is 80 µ  . The corresponding contributions to the plate height
`from the injection plug length and detector observation length
`are 0.41 µ   and 24 nm, respectively. These contributions are
`small relative to the measured plate heights. The contribution
`from axial diffusion decreases with increasing field strength.
`
`(19) Bishop, J. A Anal Chetn. Acta 1973, 63, 305.
`(20) Sternberg, J. C. Adv. Chromatogr. 1966, 2, 205.
`(21) Giddings, J. C. Dynamics of Chromatography, Parti: Principles and Theory,
`Marcel Dekker: New York, 1965; Chapter 2.
`(22) Knox, J. H.; Grant, I. H. Chromatographia 1987, 24, 135.
`(23) dc usually corresponds to the capillary diameter. In this case, a channel
`with a trapezoidal cross section is under consideration, and the channel
`depth will be used for estimating the contribution of Joule heating to the
`plate height.
`(24) Monnig, C. A; Jorgenson, J. W. Anal. Chem. 1991, 63, 802.
`(25) Hjertén, S.; Elenbring, K.; Kilár, F.; Liao, J.-L; Chen, A; Siebert, C. J.; Zhu,
`M.-D. /. Chromatogr. 1987, 403, 47.
`(26) Aebersold, R.; Morrison, H. D./. Chromatogr. 1990, 516, 79.
`(27) Jandik, P.; Jones, W. R./. Chromatogr. 1991, 546, 431.
`
`Agilent Exhibit 1282
`Page 4 of 5
`
`

`

`heating in eq 3, this contribution can be estimated. For/t(Zn) =
`= 870 V/cm, dc = 7.6 µ  ,   = 0.015
`1.65 x 10"4 cm2 V"1 s"1,
`m2 mol-1  "1, C = 80 mM, Dm =
`5 x 10"6 cm2 s"1, and  
`= 0.6
`W nr1 K-1; the contribution of Joule heating to the plate height
`is 6.5 pm. As calculated, this contribution is insignificant relative
`to the plate height data. An alternative explanation for this
`discrepancy in the experimental and calculated plate height might
`be due to geometrical fluctuations in the channel over the length
`of the separation channel. As fabrication techniques for glass and
`quartz substrates improve, this discrepancy is likely to decrease
`as was the case for silica capillaries.
`In conclusion, metal ions complexed with 8-hydroxyquinoline-
`5-sulfonic acid are resolved in less than 15 s using microchip
`electrophoresis, and detection limits are in the range of 30-60
`ppb. To enhance the detection limits for these complexes, a
`preconcentration step can be added prior to the separation. A
`sample preconcentration method that can be easily incorporated
`with electrophoretic analyses is field-amplified sample injection
`and sample stacking.25"28 Enhancements in detection limits up
`to 103 times have been observed using stacking methods with
`conventional capillary electrophoresis.29 Similar stacking proce-
`dures have been demonstrated on microchips30 and may allow
`sub-ppb detection limits for metal ions.
`
`ACKNOWLEDGMENT
`This research is sponsored by the U.S. Department of Energy
`(DOE), Office of Research and Development Oak Ridge National
`Laboratory is managed by Martin Marietta Energy Systems, Inc.
`for the U.S. Department of Energy under Contract DE-AC05-
`this research is sponsored in part by an
`840R21400. Also,
`appointment for AW-M. to the ORNL Postdoctoral Research
`Associates Program. These postdoctoral programs are adminis-
`tered by the Oak Ridge Institute for Science and Education and
`ORNL.
`
`Received for review December 13, 1994. Accepted April
`10, 1995.
`AC9412037
`
`Abstract published in Advance ACS Abstracts, June 1, 1995.
`
`3O5r
`
`e
`
` 
`
`re3
`ore9
`
`3
`
`2
`
`1
`
`O
`
`separation field strength [V/cm]
`Figure 6. Variation of plate height for Zn (O), Cd ( ), and Al (O)
`and power/length ( ) with the separation field strength complexed
`with HQS. The solid line is the calculated plate height for Zn using
`eq 3. Error bars are ±a for three analyses.
`
`However, the experimental plate height data do not asymptotically
`approach the sum of the time-independent contributions at high
`In Figure 6, the calculated plate height for Zn
`field strengths.
`using eq 3 (µ   = 1.65 x 10“4 cm2 V-1 s-1, Dm = 5 x 10"6 cm2
`s"1) is plotted. For separation field strengths of <430 V/cm, the
`experimental data correspond well with the calculated values, but
`for separation field strengths of >580 V/cm, the agreement is
`poor. This difference could be due to heating of the microchip,
`as observed in the nonlinear behavior of the linear velocities
`(Figure 4). The power/length in the separation column is also
`plotted in Figure 6 for the corresponding field strengths. Typically
`for electrophoretic systems, when the power/length remains
`interfere with the
`below 1 W/m,24 Joule heating does not
`separation performance. For separation field strengths of >580
`V/cm, the power/length is >1 W/m. Using the last term for Joule
`
`(28) Chien, R-L; Burgi, D. SChromatogr. 1991, 559, 141.
`(29) Chien, R-L; Burgi, D. S, J. Chromatogr. 1991, 559, 153.
`(30) Jacobson, S. C.; Ramsey, J. M. Electrophoresis, in press.
`
`Analytical Chemistry, Vol. 67, No. 13, July 1, 1995
`
`2063
`
`Agilent Exhibit 1282
`Page 5 of 5
`
`