RECEIVEDfor review September 16, 1974. Accepted November 4, 1974. This work was performed under the auspices of the U S . Atomic Energy Commission. Reference to a company or product name does not imply approval or
recommendation of the product by the University of California or the U S . Atomic Energy Commission to the exclusion of others that may be suitable.
An Improved Glassy Carbon Electrode Samuel C. Levy Exploratory Batteries Division, Sandia Laboratories, Albuquerque, N.M. 87 1 15
Patrick R. Farina' Scientific Glass Laboratory, Sandia Laboratories, Albuquerque, N.M. 87 1 15
A standard method of preparing working electrodes for voltammetric studies is to seal a rod of the desired electrode material in an inert sheath, cut the end to expose a flat circular cross-section of the electrode material, and then polish to reduce the surface roughness. This technique has been reported for the fabrication of glassy carbon electrodes in which the glassy carbon rod is sealed in Pyrex glass (1) and boron nitride ( 2 ) .When preparing electrodes in this manner, using Pyrex, we found that a good seal could not be obtained between the glass and the glassy carbon since the glass will not wet carbon. This resulted in a thin void between the glass and carbon into which glassy carbon dust, formed during the final polishing step, may accumulate and which fills with electrolyte during use, giving rise to anomalously high residual currents and making analysis of the data extremely difficult. To solve this problem we developed a technique for fabricating glassy carbon electrodes having a leak-tight seal. These electrodes are ideally suited for voltammetric studies in fused salts, having a small residual current and the capability to operate over a wide voltage range. Carbon-to-Glass Seal. To obtain a good seal between glassy carbon'and glass, it is necessary first to coat the carbon with a thin layer of silicon (0.0005 to 0.005 in.) ( 3 ) .Before coating, the glassy carbon rod is centerless ground to obtain the desired diameter and to remove ridges, formed during the molding of the rod, which run the length of the rod. The glassy carbon is next placed inside a Pyrex vessel and connected across the output of an auto transformer (Figure 1).A stream of silane (SiHJ, diluted with hydrogen and argon, is passed through the vessel. The variable auto transformer is turned on for approximately 2-3 seconds (42 volts, 60-cycle ac), heating the rod. Silane, upon contacting the hot carbon, decomposes and deposits a thin film of silicon on the rod. Initially we attempted to seal a silicon-coated glassy carbon rod in Pyrex. A good electrode could not be prepared because of the mismatch in expansion between the glassy carbon and Pyrex below the set point of the glass. This resulted in spontaneous cracking of the Pyrex upon cooling. A survey of the literature indicated that GSC-4, a borosilicate glass manufactured by General Electric Co., matches the glassy carbon in expansion more closely in the critical temperature range than does Pyrex (Table I). The GSC-4 glass we obtained had many bubbles in it. The presence of bubbles resulted in a poor electrode, since some bubbles invariably were located a t the glass-carbon interface. To eliminate this problem, we remelted the glass Present address, Chemistry Department, Syracuse University, Syracuse, N.Y. 604
ANALYTICAL CHEMISTRY, VOL. 47, NO. 3, MARCH 1975
-Kulgrid Leads UJ.@Oin.)
Kovar Foil Cap over C a m n Ro
4
,Flexible lead coiled to form spring for eiectrital contact pressure. Argon -100 cciminute from flowmeter
3% SiH4 in hydrogen u p t o
0.25 in. Diam. Carhon Rod
100 cciminute from flowmeter
'15-mm
OD Pyrex, 15-in. overall length
-Indent at 120' to center carbon rod (2 places on tube) into hole in urbon rod
- Nickel or Kovar 0.25 in.
diam. dish spotwelded to lead wire for electrical contact.
light burner.
Join top and bottom leads to 30-amp auto transformer: 42 volts required.
Figure 1. Apparatus for deposition of silicon on glassy carbon
at 1700 "C in a platinum crucible for 40-60 hours. This produced a glass free of crystalline inclusion and bubbles. The molten glass was cast into a rectangular block and allowed to cool. The block was core-drilled to form a rod of GSC-4 glass which was then heated and pulled down to a diameter of 4 mm. This was then made into tubing by coiling on the end of a 0.5-inch diameter Pyrex tube, remelting, and smoothing. Another Pyrex tube was attached to the end of the coiled GSC-4 glass, so it could be drawn down to a 1-mm wall thickness having an inner diameter very slightly larger than the silicon coated glassy carbon rod (Figure 2). This tubing was then slipped over the glassy carbon and sealed to it by evenly heating to the softening point in a glass lathe (Figure 3). To avoid the arduous task of remelting and forming tubes with the GSC-4 glass, one should examine the glass under a 30 power microscope and select only those portions ~~
~
Table I. Coefficients of Thermal Expansion i v e r a q * r a l u e of a ,
Matenal
renip. ranqr, 'C
Glassy carbon
0-1 00
Pyrex GSC -4
100-1000 Ck300 0-300
i mf
Crn,
O C
2.2 x 10-6 3.2 x IOm6 3.2 x 2.4 x
t
-m
-t 0. I
I 0.1
OS
,
a>
01 .
I
/
0 3 112
,
GI
,
,
, , , , , , , /
0 Q I -0.1 + 3 3 4 -05 4 6 U7 Q 8
+ 9 -1.0
" o m MRIUIAglApcI itlrn, Figure 5. Base line obtained in LICI-KCi at 500 'C ( A ) Glassy c a r ~ o nin v e x . GSC-4.0.082 Vlssc
WJO
VISBC. (a)NBW
. . . . . glassy caraon eiectrooe m
Figure 2. Formation 01 GSC-4 tubing 3w
(A) GSC-4 coiled on end of Pyrex tube, (€4 GSG4 SmOOmed and Pyrex at. tached to Other end lor drawing, (0 GSC-4 drawn to 1-mm wall thickness
210
m 1% IW "a
I /
Io0
-Io -1W -00
tive hysteresis effect. The reduction of Cr(II1) in molten LiCl-KCl has been studied using an electrode of this tvue. The results have men repor~eueisewnere 1'1.
Figure 4. Fabrication of electrode (A) Glassy carbon rod, (€4 Glassy carbon rod wilh GSC-4 Sealed to it, before inserting in end of graphite rod. (0Glassy carbon rod inserted into end of graphite rod, (0 Finished electrode. Pyrex slipped over graphite and sealed to GSC-4. End has been cut and polished
having no huhhles. This should be satisfactory for making a good seal to glassy carbon. Electrode Fabrication. Electrodes were fabricated from 1-inch lengths of glassy carbon rod, %-inch in diameter. The rods were centerless ground to a 0.120-inch diameter circular cross-section. After coating the glassy carbon with a thin film of silicon, a ?&inch length of GSC-4 glass tube was slipped over the rod and sealed to it. A hole, the same diameter as the coated glassy carbon rod, was drilled into the center of the end of a spectroscopic grade graphite rod and the end of the glassy carbon rod protruding from the glass seal was force-fit into the hole. Next, a Pyrex sleeve was slipped over the graphite rod and sealed to the GSC-4 glass. The end was cut, exposing a flat, circular cross-section of glassy carbon, which was then polished to the desired finish (Figure 4).
ACKNOWLEDGMENT The authors thank Floyd Philgreen of the Sandia Lahoratories Facilities Engineering Division for setting up the apparatus and coating the glassy carbon rods with silicon. The authors also are grateful to Charles DeMoss of the Sandia Laboratories Scientific Glass Laboratory for making the glass-to-glassy carbon seals. LITERATURE CITED (1) W . K . B e h l , J . € ~ , ~ ~ m . S o c , 118,889(1971). (2) R. Tunold, H. M. 00. K. A. Padsen. and J. 0. Ynredal. Electrochim. Acta, 16, 2101 (1971). 16. (3) Beckwith Carbon Corp.. Technical Leaflet. (3) (4) S. C. Levy and F. W. Reinhardt, Paper No. 333, 145th Electrochemical Society Meetmg. Meeting. Sa" F~IICISCO, Francisco, Calif., May 1974.
RECEIVEDfor review September 9,1974. Accepted Novemher 7, 1974. Work supported by the United States Atomic Energy Commission. ANALYTICAL CHEMISTRY. VOL. 47, NO.
3, MARCH 1975
605
