rather than a two-terminal one as used in the Jones 13ridgeJ therefore, minimizing lead resistance errors; the two comparison arms of the bridge are immersed in a constant temperature bath to eliminate heating and temperature errori; this bridge is capable of greater llrecihion than the Jones I3ridge because of the high resolution ( 0 . 0 0 0 0 2 ~ ~of) the inductive divider used in i t ; the circuit may be readily modified to make high precision capacitance and inductance measurements: and it is considerably less expensive than a Jones Bridge. EXPERIMENTAL
I n operation, a resistor of similar value to the resistance of the cell is placed in the circuit between points d and B (Figure 1) and the cell is connected between points C and D . Leads
from the oscillator and leads from the divider are connected directly to the points A and D . A balance is obtained by adjusting the divider and the capacitors for minimum response of the null detector. I n most of our studies the oscillator frequency was 1000 cycles/ second; however, any frequency from 50 cyclesjsecond to 10,000 cycles/'second may be used. For measurements of the highest precision, the resistance of the lead between B and C must be known. This lead resistance may be evaluated by replacing the conductivity cell with a resistor of known value; the circuit is balanced with the detector connected to the point B ; the detector lead is then moved to the point C and the bridge is balanced again. The difference in the two ratio measurements multiplied by the total resistance between A and D gives the resistance of the link between B and C. T o verify the accuracy of this bridge,
a calibrated resistor was substituted for the conductivity cell. Agreement between values of the resistor ratios measured at the Xational Bureau of Standards Laboratory and those measured in our laboratory was *0.2 p.p.m, ACKNOWLEDGMENT
The authors acknowledge the cooperation and guidance of Thomas L. Zapf and Patrick H. Lowrie, Jr., of the National Bureau of Standards, and Claude R. Daum and Garth Ghering of the U. S. Geological Survey. LITERATURE CITED
(1) Jones, G., Josephs, R. C., J . Am.
Chem. SOC.50, 1949 (1928). (2) Thomas, J. L., KBS Circular 470 (1948). PUBLICATIOK authorized by the Director, U. S. Geological Survey, Washington, D. C.
A Time Base Generator with Digital Readout of Nuclear Magnetic Resonance Line Positions J. M. Purcell and J. A. Connelly, Eastern Utilization Research and Development Division, U. S. Department of Agriculture, Philadelphia, Pa. 191 18 a n X-Y recorder as a T readout device with Varian HR-60 high resolution nuclear magnetic resHE USE OF
onance (NMR) spectrometers is well established. I n this application the X-axis is driven by an external time base. Spectral calibration is normally accomplished by the sideband technique ( I , 8 ) . Line positions are measured by first obtaining a scale factor between the frequency scale and linear distance. The linear distance is measured from each N M R line to the reference line. The value of the linear distance multiplied by the scale factor yields the line position in terms of frequency (cycles per second), Several disadvantages are inherent in this method of measuring N M R line positions, not the least of which is the amount of time involved. The chief characteristic which affects the accuracy of the NMR line positions is the linearity of the external time base and! or of the recorder. Commercial recorders generally possess linearity of the order of 0.1 to 0.2% of full scale. If the linearity of the external time base is greater than the recorder linearity (considering only the readout device), the accuracy of the line positions is limited only by the linearity of the recorder. The inherent nonlinearity of the slidewire accounts at least in part for the nonlinearity of the recorder. The device described herein provides a more accurate and convenient method for measuring N M R line positions
through the use of an operational amplifier in two modes of operation. The accuracy of N M R line position measurements can be increased in two ways: improving time base linearity and avoiding slidewire nonlinearity. Consider first time base linearity which may be defined as constancy in the time rate of change of a ramp voltage. An excellent method for the production of a linear ramp voltage is the integration of a constant voltage. Assuming in principle that this method will yield a precisely linear ramp voltage, one can see that in pract,ice the linearity of the time base depends upon the specifications of the electronic devices used. We have designed a time base generator of improved linearity using a solid state voltage source with a low temperature coefficient and a high gain, chopperstabilized operational amplifier. Consider now the inherent nonlinearity of recorder slidewires. S o n linearity is the result of the fact that the resistance of t'he wire per unit, distance is not constant. Therefore, when one attempts to correlate voltage with linear distance, using another scale which possesses a linearity different from that of the slidewire, the result is a certain unavoidable error. This error can be avoided by driving the, pen through the same distance along the same portion of slidewire by means of an accurate voltage source and then correlat'ing that voltage with distance. The necessit.y of measuring distance on
the chart paper is thereby eliminated, since for a given recorder range the voltage necessary to reach a given point on the slidewire is now known. This voltage can be correlated with the quantity plotted along the abscissa. However, to achieve the greatest possible accuracy in measuring line positions the measurement should be made before the chart has been removed from the recorder. By employing this principle we have devised a method which provides a fast digital readout of N M R line positions in terms of frequency. EXPERIMENTAL
The device consists of a combination of two modes of operation of an operational amplifier (Dymec Model DY2460A-M1 with Model DY-2461A-bI5 blank plug-in; 2-p.p.m. d.c. gain accuracy including linearity)-i.e., integration of a constant voltage and +1 gain amplification. A schematic representation is given in Figure 1. A photographic view of the time base generator constructed in this latoratory is given in Figure 2. For the integration mode the time constant is 100 seconds. Since stability of the time constant is important, a 10-megohm metal film resistor and a 10-pf. polystyrene capacitor were used. Proper connections for the Dymec 2460A operational amplifier can be found in the Dymec instruction manual. Ramp voltage linearity depends upon the voltage source stability. T o proVOL. 37, NO. 9, AUGUST 1965
*
1 1 81
r
0.1%
1/2 w
Voltage
Divider lOKn
.
LO.
output
L -
-
-
.Hi
Figure 1. Circuit diagram for time b a r e generator with digitol reodout of NMR line positions vide the greatest stability a solid state voltage reference source (CircuitDyne Corp. Model PRIB-115-11.0 G ) with a temperature coefficient of i0.00170 was chosen. The 11.061-volt output of the voltage source is reduced by means of a 40K ohm resistor and a 10K Trimpot in series with a 10K ohm, fourdecade voltage divider (General Resistance Co. Dial-A-Vider Model DV4004; linearity a t one watt is +40 p.p.m.). Adjustment of the voltage across the decade divider to 2.0000 volts is accomplished by the Trimpot. The output of the voltage divider is variable in decade fashion from O.Ml00 to +2.0000 volts. With the selection of this voltage a variation of pen speeds from a maximum of 5 seconds per inch to a minimum of 2.7 days per inch can he achieved. Immediately following the voltage source a D P D T (ON-OFFON) switch is employed as a polarity reversing switch so that the recorder pen may he driven in either direction or may he stopped in any position. The right to left direction, which is not included on commercial recorders, is more convenient for the recording of N M R spectra. Convenience results because the recording of the spectrum can commence at the reference line (usually tetramethylsilane) at the righthand side of the chart. The spectrum is displayed according to the usual convention, except that the ringing will
be on the low field side of the various lines. The conversion from one mode of operation to the other is accomplished hy means of a four-pole, double-throw (OS-ON) switch. One advantage of this type of time base over one produced by a motor driven potentiometer is that the former permits rapid return of the pen to zero position. This is accomplished by a normally-open push button switch which short circuits the capacitor. A banana jack is connected across the output of the voltage reference supply and is mounted on the front panel as a convenient reference voltage source for other purposes, such as calibrating other instruments. DISCUSSION
For the time base mode (with the four pole, double throw (ON-ON) switch in the proper position) the pen speed is determined by the settings of the voltage divider and the recorder X-axis range. The direction of pen travel is determined by the position of the D P D T (OS-OFF-ON) switch. Control of whether the pen touches the paper is retained in the recorder so that, if desired, the pen may be driven along the X-axis without touching the paper.
Figure 2. 11 82
ANALYTICAL CHEMISTSY
Photographic view of front panel
N M R line position measurements are made in the following manner: The chart paper with the spectrum in placed on the recorder platen, using the vacuum hold-down. With the pen up, the decade divider set at O.Oo(10 and the D P D T (ON-OFF-ON) switrh in either O S position the point of the pen is visually aligned with the center of the reference peak. Then the side hand frequency is dialed on the decade divider-e.g., 624.8 c.P.s., the decimal point occurring between the third and fourth decades. The pen is then visually aligned with the center of the side-hand peak by adjusting the r e corder X-axis range control and vernier. Finally, the decade divider is dialed until each N M R line in turn is visually aligned with the pen point. The reading on the divider is the line position in terms of frequency (c.p.s.). The time base output has a linearity of O . O Z ~ oof full scale or better. As the voltage accumulated at a given voltage divider setting time measurements were made at incrementa of 1 volt. Two timevoltage pairs were used to determine the equation of a straight line. Voltages corresponding to the remaining time measurements were calculated and compared with the observed values. The reproducibility is 0.03Y0 of full scale or better. For a given setting of the voltage divider, the accumulated output ramp voltage both negative and positive at 1000 seconds was measured to the nearest 0.1 my. The reproducibility figure was calculated by dividing twice the standard deviation of the voltage measurements hy the mean value of the output ramp voltage. The line position (+1 gain) output has a linearity of 0,007570 or better. The voltage corresponding to each digit of the first decade of the voltage divider was measured. Two divider settingvoltage pairs were used to determine a straight line. Voltages corresponding to the remaining divider settings were calculated and compared with the o b served values. A comparison of the conventional and the electronic (described above) methods of measuring N M R line positions is shown in Table I, which gives two average values of the line position in terms of c.p.s. from tetramethylsilane
and the corresponding standard deviations for each line. Ten recordings of each spectrum were obtained at room temperature, using a Varian DP-60 spectrometer. T o avoid signal saturation the radio field amplitude was set 60 db, below one-half watt. The external magnetic field was swept at the rate of 0.5 to 1.0 c.p.s./second. Spinning 5-mm. 0.d. sample tubes were used to achieve maximum field homogeneity. Calibration of the spectra was accomplished by the usual side-band technique, using a Hewlett-Packard 200 C D wide band oscillator and a 522B counter. Tetramethylsilane (K & K Laboratories Inc., Jamaica, N. Y.) was used as an internal reference. Several advantages attend the use of the time base generator described above. I n addition to improved linearity and reproducibility, the time base generator affords greater convenience in the recording of IVMR spectra, since the pen may be driven in the right-to-left direction. By convention IVMR spectra are displayed with magnetic field strength increasing from left to right. Using tetramethylsilane as an internal reference, the reference signal normally occurs further upfield than all other spectral lines. Therefore, one can begin
Table 1.
comparison of Two Methods of Measuring NMR Line Positions
Conventional method Line" position Std. dev.b 1,1,2-Trichloroethylene Chloroform Methylene of ester group of ethylcyanoacetate Methanol
387,4 434.5 243 3 250,3 257.7 264.8 197,8 202,8 252.9 258.0 263 0 3 216 8
ass
Methoxy group of methyl butyrate C.P.S.from tetramethylsilane, solvent-CClr * Two sigma values.
0.32 1,44 1.25 1.20 1.17 1.14 0.73 0.76 0.78 0 89 0.89 0.88 1.32
Electronic method Linea Std. dev. position 0.21 0.30 0.50 0.64 0.60 0.57 0.70 0 . 7.5 0,77 0.70 0.75 0.75 0 70
387.3 434.8 243.4 290. 6 257 8 264 9 197 8 202 9 252 7 258 0 263 2 268,2 216,2
0
a t the beginning with only an irrelevant change in the appearance of the spectrum. The possibility of error due to the nonlinearity of the recorder slidewire has been eliminated. Consequently, line position measurements are more precise. Convenience also is achieved by correlating voltage with linear distance along the recorder slidewire. Linear distance on the chart paper need
not be measured. Also, line positions are rapidly measured in digital form in terms of frequency. LITERATURE CITED
(1) Arnold, J. T., Dissertation, Stanford University (19.54). (2) Arnold, J. T . , Packard, hI. E., J. Chem. Phvs. 19, 1608 (1951). *Mention of commercial products does not
constitute an endorsement by the I!. S. Department of Agriculture over others of a similar nature not mentioned.
Combining Gas Chromatography with Nuclear Magnetic Resonance Spectrometry E. G. Brame, Jr.,' Manufacturing Division, Plastics Department, E. I. du Pont de Nemours & Co., Wilmington, Del. has been compreviously with nuclear magnetic resonance spectrometry (NMR) for the identification of effluent fractions from Ce aldehydes ( 4 ) , isoprenoid hydrocarbons (2), isomers of triisobutylene and tetraisobutylene (3) and C, olefins (1). I n each of these cases conventional trapping procedures were used. They consist of trapping samples into one vessel and transferring them into another vessel for the S M R examination. These procedures require a relatively large amount of starting material and generally can take a long time (hours or more) to collect a sufficient amount of pure sample for identification. I n the present investigation, a new procedure which eliminated the use of a second vessel was developed. Thus, the same vessel into which the sample was trapped could also be used for the N M R examination. .4s a reAS CHROMATOGRAPHY
G bined
1 Present address, Elastomer Chemicals Department, E. I. du Pont de Nemours & Company, Wilmington, Del., 19898.
sult of this procedure less sample was required to obtain N M R spectra. The new procedure of combining gas chromatography with N M R was developed with the use of glass microt'ubes (NMR Specialties, Inc., New Kensington, Pa.) that became commercially available only last year for N M R studies. These tubes are made of. borosilicate glass and contain a bubble a t about 15 mm. from the bottom of the tube to coincide with the position of the receiver coil in the spectrometer. The exact position can be determined by filling the bubble with chloroform and by noting the position that corresponds to the maximum signal for the proton resonance. X l-mm. i.d. capillary leads from the top of the tube to the bubble whose volume is about 35 pl. After several attempts were made t o trap efluents directly from a gas chromatograph into these tubes using a 22-gauge syringe needle that was eight inches long, it) was found that this fine gauge needle caused a very large pressure drop a t the esit of the gas chromatograph. Inefficient operation of the gas chromatograph resulted. T o circumvent this difficulty, a pair
19898
of microtubes t h a t had a 2-mm. i.d. capillary leading to the bubble were obtained on special request. With the larger capillary, a larger diameter syringe needle (17-gauge) could be ueed for trapping samples. Thus, less restriction to the helium flow and greater efficiency in trapping was made possible. ii picture of the microtube and equipment used for trapping samples directly from a gas chromatograph is shown in Figure 1. The 17-gauge needle that is eight, inches long is seen inside the glass microtube. The tip of the needle is located inside the bubble near the bottom of the tube. The top of the needle is connected to a Luer lock which is silver soldered to a inch tubing T. I3y using a Luer lock connection, the needle can either be attached or removed very quickly and easily from the tubing T. The right side of the T is equipped to connect it to the exit port of a gas chromatograph. The left side is shown uncapped. Rhenever a species to be trapped emerges from the esit port, the nut to the immediate left of the T is screwed on the T and tightened to force the carrier gas to pass through the needle in the VOL. 37, NO. 9 , AUGUST 1965
1 1 83
