lts for ""
Below is a cache of http://www.uvp.com/pdf/wavelengths.pdf. It's a snapshot of the page taken as our search engine crawled the Web.
The web site itself may have changed. You can check the current page or check for previous versions at the Internet Archive. Yahoo! is not affiliated with the authors of this page or responsible for its content.
Wavelengths of spectral lines in mercury pencil lamps a reprint from Applied Optics
Wavelengths of spectral lines in
mercury pencil lamps
Craig J. Sansonetti, Marc L. Salit, and Joseph Reader
The wavelengths of 19 spectral lines in the region 253-579 nm emitted by Hg pencil-type lamps were measured by
Fourier-transform spectroscopy. Precise calibration of the spectra was obtained with wavelengths of
198
Hg as
external standards. Our recommended values should be useful as wavelength calibration standards for
moderate-resolution spectrometers at an uncertainty level of 0. 000 1 nm.
Key words: Mercury pencil lamp, wavelengths, Fourier-transform spectroscopy, spectral lines.
1. Introduction
Hg pencil-type discharge lamps are widely used for
alignment and calibration of instruments in analytical
spectroscopy laboratories. Their wide availability, easy
operation, and simple spectrum make them an attractive
source for wavelength calibration; however, no precise
measurements of the Hg lines as emitted by these lamps have
previously been published. As one component of a
Cooperative Research and Development Agreement between
the National Institute of Standards and Technology and Oriel
Instruments, we have observed several Hg pencil lamps with
a high-resolution Fourier-transform spectrometer (FTS) and
have evaluated their suitability as a source of wavelength
standards. The lamps have also been evaluated for their
usefulness as radiometric standards. Results of the
radiometric measurements are presented in an accompanying
paper.
1
2. Experiment
All samples of the Hg pencil lamp used in this work were
supplied by Oriel Instruments.
2
The lamp (Oriel Model 6035)
consists of a quartz tube with U-shaped capillary filled with
Ar at a low pressure and a few milligrams of metallic Hg of
natural isotopic composition (Fig. 1). The lamp operates as a
low-current discharge with either ac or dc excitation. For our
measurements the power supply (Oriel Model 6060) was
used in dc mode with a discharge current of 15.00 ± 0.01 mA
as the standard operating condition. The lamp was mounted
unshielded
The authors are with the National Institute of Standards and Technology,
Physics A167, Gaithersburg, Maryland 20899.
Received 14 April 1995; revised manuscript received 7 August 1995.
in a vertical orientation with the base down and operated in
ambient air at a temperature of approximately 23
o
C. The air
surrounding the lamp was not subject to drafts. The lamp
was rotated about a vertical axis so that the optical axis of
the FTS passed through both legs of the U-shaped
capillary.
All measurements were made with a Chelsea FT-500
Fourier-transform spectrometer.
2,3
This instrument is
optimized for response in the near-ultraviolet region and is
capable of operation from approximately 180 to 900 nm.
Each Hg lamp was observed in two overlapping
wavelength regions: 230-460 and 375-600 nm. The spectral
bandpass for each region was limited by optical filters and
by the responses of the photomultiplier tubes used as
detectors. All our spectra were recorded at a resolution of
0.03 cm
-1
(0.0002 nm at 250 nm to 0.0010 nm at 580 nm),
which is the maximum resolution of the instrument. The
FTS was operated in a dual channel mode, recording as the
interferogram the difference of the two detector signals. For
most of our spectra three to five interferograms were co-
added to improve the signal-to-noise ratio. Typical values
of the signal-to-noise ratio ranged from ~10
4
for the
435.8-nm line to ~40 for the weak line at 434.7 nm.
To obtain absolute calibration of the pencil-lamp spectra,
precisely known lines of
198
Hg were used as external
wavelength standards. Each observation of a natural-Hg
pencil was preceded and followed by the observation of
198
Hg lamp made under identical conditions.
198
Hg was
excited in an electrodeless discharge lamp containing Ar at
a pressure of 33 Pa (0.25 Torr). The lamp was operated
with ~70 W of microwave power over an open-cup antenna
and was cooled with a gentle stream of air from a fan.
Values for the standard wavelengths were taken from
Kaufman.
4
74 APPLIED OPTICS / Vol. 35, No. 1 / 1 January 1996 To obtain an accurate calibration of Fourier spectra with
an external standard source, it is necessary that the unknown
and the standard sources illuminate the instrument in the
same way. This problem has been discussed in some detail
by Learner and Thorne.
5
Our
198
Hg discharge was contained
in a quartz capillary of approximately the same diameter as
the pencil lamp. Both sources were placed directly in front of
the aperture of the FTS, with no external optics, at a distance
chosen so that the source just filled the collimating mirror of
the instrument. We aligned the sources on the axis of the
instrument by opening the input aperture to its maximum
size, observing the source through one of the detector
apertures of the FTS, and visually centering the source in the
aperture. The reproducibility of the results obtained when the
198
Hg source was removed and realigned demonstrate that
this was a satisfactory alignment procedure at our desired
level of accuracy.
For comparison with the pencil lamps, and as a general
test of the accuracy of our measurement methods, we also
made a single observation of the spectrum of natural Hg
excited in an electrodeless discharge lamp. The lamp
consisted of a quartz tube of 10-mm inner diameter
containing a small quantity of natural Hg and Ar at a
pressure of 400 Pa 3 Torr). It was operated with ~60 W of
microwave power over an open-cup antenna with gentle air
cooling.
3. Analysis
Because natural Hg contains six isotopes in substantial
abundance, two of which have a magnetic hyperfine
structure, all the Hg lines show complex line profiles at high
resolution, as illustrated in Fig. 2. For most lines, the
transitions of the four even number isotopes are unresolved
in a strong asymmetric feature near the center of the line
profile. Components of the odd-number isotopes, which are
more widely split by hyperfine structure, appear as partially
and fully resolved satellites. The detailed appearance of each
line and the wavelength of the maximum-intensity point in
the profile are critically
Fig. 2. (a) Spectrum of the 404.6-nm line of natural Hg from a
pencil-type lamp at an instrumental resolution of 0.03 cm-1 (0.0005nm).
(b) The same spectrum with the resolution degraded to 1.0 cm-1 (0.016
nm) by convolution with a Gaussian instrumental function.
dependent on the resolution of the spectrometer with which
the line is observed. Because setting on the point of
maximum intensity is the most convenient and commonly
used method for determining the position of a spectral line,
the structures of the Hg lines are an important
consideration when they are to be used as wavelength
standards.
We tested the effect of instrument resolution on the
usefulness of these complex lines as wavelength standards
by convolving the observed high-resolution spectra with
Gaussian profiles of varying widths. From these tests we
determined that all observed lines appear as simple
symmetric profiles if the resolution is degraded to 1.0 cm
-1
(from 0.0063 nm at 250 nm to 0.034 nm at 580
nm). Any
further reduction in resolution broadens but does not shift
the lines. The results we present apply to the case in which
the resolution is low enough that the structure is entirely
unresolved. For this case the position of the line is the same
whether it is determined as the center of gravity or as the
point of maximum intensity. At higher resolutions all lines
eventually become asymmetric and display shifts in the
position of the intensity peak with changing
1 January 1996 / Vol. 35, No. 1 / APPLIED OPTICS
75 resolutions. For this reason use of our present results should
be limited to wavelength calibration of spectrometers with
resolving powers of less than ~ 17 000.
Although our data were recorded and transformed to
produce a resolution of 0.03 cm
-1
, both the pencil lamp and
the
198
Hg spectra were convolved with a 1.0-cm
-1
Gaussian
profile to degrade the resolution before any wavelength
determinations were made. The line positions were then
determined by use of a quadratically smoothed,
first-derivative algorithm to locate the intensity peak of the
profile. Tests showed that this rapid, simple procedure
produced results that differed negligibly from a more time
consuming procedure in which the entire line profile was
fitted. The measured values for the lines of
198
Hg were
compared with the literature values
4
to determine the
multiplicative correction factor to be applied to the wave
numbers. Typically the correction factor for a single
198
Hg
spectrum was defined to better