1046
ANALYTICAL CHEMISTRY, VOL. 50, NO. 8, JULY 1978
NHB, the frequency change obtained was about 386 Hz. A plot of the logarithm of change in frequency (log AF)vs. the logarithm of concentration (AC) is linear over the range of concentrations 0.01 ppb to 1 ppm. Table 11 illustrates the results of an interference study. Since the concentrations of all the interfering substances are much higher than the concentration of ammonia injected, and the responses are much less than the responses of ammonia observed, no interference is expected. The response time observed was less than 30 s using the pyridoxineSHC1 coating and a complete reversibility of response was observed in 4 min. In order to determine the type of reaction which exists between pyridoxineSHC1 and ammonia, infrared spectra were taken. Infrared spectra of pure pyridoxine-HC1 (A) and pyridoxineeHC1 exposed to ammonia (B) are shown in Figure 3. In spectrum A, the absorption band at 3250 cm-l is from VO-H stretching vibration. In spectrum B, this band disappeared and a new, broad band between 3150 and 3100 cm-’ can be seen, which indicates the presence of the NH4+group (VN-H stretching vibration in NH4+salts). The absorption band a t 1400 cm-l is from 8N-H bending vibration and provides an evidence that the reaction between the hydroxyl group and ammonia has occurred, yielding the ammonium salt of pyridoxine.HC1.
CHZOH
As the pyridoxineHC1 coating is very sensitive for ammonia, the detection of trace quantities of ammonia is possible. Consequently, if the coating substrate is exposed to a higher concentration of ammonia, the pyridoxineHC1 could be saturated and less reproducibility is obtained. Care must be taken to use this sensor only for ppt and ppb concentrations of ammonia.
LITERATURE CITED (1) (2) (3) (4) (5) (6) (7)
K. H. Karmarkar and G. G. Gullbault, Anal. Chim. Acta, 75, 111 (1975). L. M. Webber and G. G. Gullbault, Anal. Chem., 48, 2244 (1976). J. Hlavay and G. 0. Gullbault, Anal. Chem., 50, 965 (1978). G. 2. Sauerbrey, Z.Phys., 155, 206 (1959). G. 2 . Sauerbrey, 2. Phys., 178, 457 (1964). K. H. Karmarkar and 0. G. Gullbault, Anal. Chim. Acta, 71, 419 (1974). F. W. Karasek and J. M. Tiernay, J . Chromatogr., 89, 31 (1974).
RECEIVED for review December 5, 1977. Accepted April 26, 1978. The authors gratefully acknowledge the financial support of the Army Research Office, in the form of Grant No. DAAG-77-G-0226, in carrying out this research project.
Continuous Determination of Hydrogen Extractable Nitrogen from Silicon-Iron by an Ammonia Gas Sensing Electrode R. G. Hirst” and C. M. Mauclone Power Delivery Technical Resources Operation, General Electric Company, 100 Woodla wn Avenue, Pittsfield, Massachusetts 0 120 1
A technique has been developed to continuously monitor the nltrogen egress from bulk samples of 3 % Si-Fe annealed in hydrogen. The apparatus and technique are applicable to variable heating rates, isothermal condiiloning, and other metal systems. By comparlson with published data on nitride solubilitles and by selectlng materlais of dlfferent chemistry] peak assignments have been made for “Soluble” nitrogen] a-Si3N4,AIN, and BN. These specles help define the nature and role of nitrogen bearing species In the secondary recrystallization of sllicon-iron. Examples are shown in application of this technique to generically different silicon-iron composlt Ions.
In silicon-iron processing, many agents, which are reported to promote selective grain growth, have been studied. These include aluminum nitride, vanadium nitride, manganese sulfide, silicon nitride, solute sulfur, and solute boron. In addition to the above agents, nitrogen is a controlled element and is present in most commercial melts at significant levels (30 to 100 ppm). This nitrogen has been shown to play an active role in the secondary recrystallization phenomenon ( I ) . Most studies of nitrogen in silicon-iron involve chemical dissolution of the matrix and separation of an insoluble residue. The fraction of the species dissolved is termed the soluble portion and the undissolved fraction is considered the insoluble portion. Soluble and insoluble in this context refer to chemical, not metallurgical solubility. Metallurgically soluble or “hydrogen extractable” nitrogen is believed to be 0003-2700/78/0350-1046$01 .OO/O
more significant in understanding secondary recrystallization. A method has been reported (2, 3) for the determination of “hydrogen extractable nitrogen” in low alloy steels by isothermal annealing at 500 “C. The ammonia formed by the extraction of nitrogen from the steel by hydrogen is absorbed in dilute acid and the ammonium measured spectrophotometrically. Headridge and Long ( 4 ) improved on the method by substituting an ammonium ion sensitive electrode to continuously measure the ammonia captured in an aqueous solution. While Headridge and Long succeeded in continuously monitoring a solution for ammonium ion, they measured a cumulative or integrated ammonium concentration and not a real time concentration in the hydrogen carrier gas. The technique reported herein utilizes an ammonia gas-sensing electrode to measure the dynamic evolution of nitrogen (as a function of temperature) from a bulk sample directly in the gas phase.
EXPERIMENTAL A system is provided which heats a sample at a known and controllable rate while simultaneously allowing that sample to interact with the surrounding atmosphere. The heating chamber consists of a tube furnace capable of continuous operation at temperatures up to 1250 O C with a thermal controller able to maintain a top soak temperature of *2 O C . A 30-inch long quartz (2.5-cm id.) tube serves as the sample retort. Stainless steel caps fit over the ends of the quartz tube and are bolted against rubber O-rings providing for a tight gas seal (Figure 1). The incoming hydrogen gas passes through an anhydrone drying tower, a regulating valve, a calibrated flow 0 1978 American Chemical Society
A ARGON 8 HYDROGEN C CONTROL VALUE D MAGNESIUM PERCHOLORATE DRYING TOWER E FLOWRATE F ENDCAPS G SAMPLE MANIPULATOR H QUARTZ RETORT I SAMPLE J FURNACE K COLDFINGER L PLASTIC BLADDER M GAS SENSING ELECTRODE
RECORDER METER
TEMPERATURE
RECORDER
CONTROLLER
Figure 1. Schematic of apparatus for nitrogen determination
Table I. Quantitative Comparisons Alloy A B C
D
Atomic nitrogen removed during heat treatment in hydrogen
Nitrogen remaining after heat treatment in hydrogen
Total
Bulk nitrogen prior to heat treatment in hydrogen
23 36 20 22
33 26 35 39
56 62 55 61
68 68 69 69
meter, and through a port in the metal cap at the entrance end. This cap is provided with a second port to house a hooked stainless steel rod for sample manipulation. Gas exits through a port in the exit cap, passes through a cold finger (cold water) and into a fritted glass bubbling tube housed in a plastic chamber. The gas exits from the top of the chamber through a back pressure bubbler. The log concentration of the ammonia gas is measured using an ammonia gas-sensing electrode inserted through the top of the chamber and placed about 1 cm above the bubbler. A small amount of water is placed in the chamber t o maintain a high relative humidity. A hole is provided through the outside edge of the exit cap for the purpose of housing a measuring thermocouple along the outside wall of the retort. The control couple i3 connected to an X-Y recorder and the electrode is connected through an expanded range millivoltmeter to the X-Y recorder. The progress of nitrogen evolution as a function of temperature is monitored instantaneously. A coupon, 3 X 5 x 0.028 cm and weighing approximately 2 g, is placed in the cold zone of the quartz chamber. The chamber is sealed and flushed approximately 15 min with argon. With the millivolt meter zeroed and the electrode in place, the gas is switched to hydrogen and the flow rate regulated. The sample is pushed into the hot zone and the sample manipulator withdrawn. After equilibration,a programmed heating rate is initiated. Since the electrode measures the log concentration of ammonia, the X-Y plot gives temperature vs. log concentration. A computer program was written to convert the traces to linear coordinates. By calibrating with a standard gas (50 ppm NH3) and knowing the gas flow (6.4 Ljh) and heating rates, this program has been extended to integrate the areas under the curves to yield a semiquantitative determination of the nitrogen evolved. A comparison of the difference between the initial nitrogen in the sample and the final nitrogen after anneal with the evolved nitrogen gives reasonable correlation (See Table I).
MOBILE NITROGEN EVOLUTION
IEMPERATURE ‘C
Figure 2. Schematic of temperatures of nitrogen evolution based on
present experimental conditions
RESULTS AND DISCUSSION The success of the method depends on the extraction of atomic nitrogen and the formation of ammonia a t the steel surface by interaction with hydrogen. As the temperature of the sample is increased and select nitrides dissolve in the matrix, “new” atomic nitrogen is created to combine with hydrogen to form ammonia. Thus a temperature excursion under specific operating conditions will yield a nitrogen “fingerprint” of the sample under study with peaks corresponding to egress maxima of “soluble” nitrogen and dissolvable nitrides. By comparison of peak temperatures with solubility data taken from the literature (5-8) and by observation of traces generated on materials of different initial chemistry, “soluble” nitrogen (i.e., nitrogen in the interstices of the matrix), asilicon nitride (a-Si3N4),aluminum nitride (AlN), and boron
1048
ANALYTICAL CHEMISTRY, VOL. 50, NO. 8, JULY 1978
Table 11. Melt Compositions, ppm Sample
Tia
Hi Zr,Ti Ferroboron Low Zr,Ti Ferroboron B / N = 2.1 BIN = 1.2
-590 40 < 50 < 50 < 50 < 50 < 50 < 50
- 90
-
B/N = 0.5
BIN= 0.2 MnS “inhibited” AlN “inhibited” a
Z i
< 50 < 50 < 50 < 50 < 50
