Development of a high-resolution laser absorption spectroscopy

Subscriber access provided by UNIV OF NEW ENGLAND ARMIDALE

Article

Development of a high-resolution laser absorption spectroscopy method with application to the determination of absolute concentration of gaseous elemental mercury in air Abneesh Srivastava, and Joseph T. Hodges Anal. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.analchem.8b00757 • Publication Date (Web): 30 Apr 2018 Downloaded from http://pubs.acs.org on May 1, 2018

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Analytical Chemistry

1

Development of a high-resolution laser absorption spectroscopy method with

2

application to the determination of absolute concentration of gaseous elemental

3

mercury in air

4

Abneesh Srivastava* and Joseph T. Hodges

5

Chemical Sciences Division, National Institute of Standards and Technology, 100 Bureau Drive, Gaithersburg, Maryland 20899-8393, United States

6

*Corresponding author: Abneesh Srivastava, [email protected]

7 8

ABSTRACT

9

Isotope dilution cold-vapor inductively coupled plasma mass spectrometry (ID-CV-ICPMS)

10

has become the primary standard for measurement of gaseous elemental mercury (GEM) mass

11

concentration. However, quantitative mass spectrometry is challenging for several reasons

12

including: (1) the need for isotopic spiking with a standard reference material, (2) the

13

requirement for bias-free passive sampling protocols, (3) the need for stable mass spectrometry

14

interface design, and (4) the time and cost involved for gas sampling, sample processing and

15

instrument calibration. Here, we introduce a high-resolution laser absorption spectroscopy

16

method that eliminates the need for sample-specific calibration standards or detailed analysis of

17

sample treatment losses. This technique involves a tunable, single-frequency laser absorption

18

spectrometer that measures isotopically resolved spectra of elemental mercury (Hg) spectra of 6

19

1

20

modeled from first principles using the Beer-Lambert law and Voigt line profiles combined with

21

literature values for line positions, line shape parameters, and the spontaneous emission Einstein

22

coefficient to obtain GEM mass concentration values. We present application of this method for

S0 ← 6 3P1 intercombination transition near λ = 253.7 nm. Measured spectra are accurately

1 ACS Paragon Plus Environment

Analytical Chemistry 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 29

23

the measurement of the equilibrium concentration of mercury vapor near room temperature.

24

Three closed systems are considered: two-phase mixtures of liquid Hg and its vapor, and binary

25

two-phase mixtures of Hg-Air and Hg-N2 near atmospheric pressure. Within the experimental

26

relative standard uncertainty (0.9-1.5) % congruent values of the equilibrium Hg vapor

27

concentration are obtained for the Hg-only, Hg-Air, Hg-N2 systems; in confirmation with

28

thermodynamic predictions. We also discuss detection limits and the potential of the present

29

technique to serve as an absolute primary standard for measurements of gas-phase mercury

30

concentration and isotopic composition.

31 32

Keywords: Mercury, Laser Absorption Spectroscopy, Standard Generator, SI traceability, Dumarey equation, vapor pressure correlation, saturation, detection limit, isotope ratio

33

INTRODUCTION

34

There are three forms1,2 of atmospheric mercury: gaseous elemental mercury, Hg0 (GEM),

35

divalent gaseous mercury (Hg(II)) and particle-bound mercury (Hg-p). The majority (95%) of

36

the gaseous mercury is in the form of Hg0, which makes the emission, transport and toxicology

37

of GEM an important species in atmospheric science studies.3 GEM emissions arise from a

38

combination of anthropogenic and natural geogenic primary sources and secondary re-emission

39

sources originating from surface reservoirs3. GEM has an atmospheric lifetime of (0.5 to 1) yr,4

40

allowing it to be transportable and prone to bio-accumulation.1,5 Mercury is classified as a

41

neurotoxin,6,7 and it is regulated by the United States Environmental Protection Agency ( US

42

EPA) and other global government agencies as a potent hazardous pollutant.8,9 The

43

environmental and occupational health standards for inhalation exposure to mercury vapor range

44

over (0.3 to 100) µg m-3, classified based on time, threshold and trigger for investigation.10


Page 3 of 29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


45

In order to quantify mercury emissions in the field, continuous emission monitoring systems

46

comprising a calibration GEM generator unit and cold-vapor atomic absorption or fluorescence

47

spectroscopy analyzers, (CVAAS, CVAFS) unit are commonly employed.11 GEM generators are

48

open systems that involve a metered carrier gas (typically air or N2) in equilibrium with a

49

reservoir of liquid Hg at a controlled temperature. In principle, this approach, provides a GEM

50

mass concentration that depends only on the partial pressure of the liquid mercury and the total

51

flow rate of the carrier gas. In practice, temperature and flow rate are used to precisely control

52

the amount of GEM in the carrier gas, and linkage to the SI (Système International d’unités) is

53

realized through a chain of calibrations to an SI-traceable primary standard for elemental

54

mercury mass concentration.

55

In the USA, primary standard generators for mercury mass concentration are used to certify

56

commercial standard generators to which field-deployed GEM systems are typically certified.

57

The primary standard generators used for these certifications, which are located at the National

58

Institute of Standards and Technology (NIST) in Gaithersburg, Maryland span the mass

59

concentration range (0.25 to 300) µg m-3 for GEM and are in turn characterized by SI-traceable

60

measurements of mercury mass and carrier gas volume. Specifically, the mass of mercury

61

emitted by the NIST primary standard generators is measured using an isotope-dilution

62

inductively coupled mass spectrometry (ID-ICPMS) method that is referenced to an SI standard

63

for mercury amount-of-substance (NIST SRM 3133).

64

NIST primary standard generators have relative standard uncertainties between 1 % and 2 % for

65

GEM mass concentration.

12

Using this scheme, certifications of

66

Although mass spectrometric measurements of collected mercury result in acceptable

67

accuracy for characterizing the output of the NIST standard generators, the overall approach is



Page 4 of 29

68

time consuming, labor intensive, and costly. Therefore, at the expense of introducing additional

69

uncertainty because of unaccounted-for system drift, routine certifications of the NIST standard

70

generators cannot be implemented. Consequently, we seek an alternative approach that is rapid,

71

accurate, and which has the potential for mercury concentration measurement over a wider

72

dynamic range than is currently available. The present strategy involves demonstrating the

73

feasibility of isotope-resolved, ultraviolet (UV) laser absorption spectroscopy of elemental

74

mercury: an SI-traceable, calibration-free method that will be sensitive, rapid and accurate.

75

Direct absorption of ground state atoms, as adopted in our approach, provides an advantage over

76

excited state-relaxation-based fluorescence detection, which is sensitive to quenching by foreign

77

gas. As discussed below, we project that this technique will measure mercury mass

78

concentrations greater than approximately 0.1 µg m-3 using single-pass absorption cells

79

nominally 1 m in length. Furthermore, combining the high-resolution approach discussed here

80

with pathlength-enhanced methods such as cavity ring-down spectroscopy (CRDS) of GEM13-16

81

will extend the accessible measurement range to ng cm-3 concentration levels, which are

82

representative of ambient-level abundances and well outside the capabilities of current standards

83

maintained at NIST.

84

In addition to developing a new method for measuring the concentration of GEM, this study

85

addresses present discrepancies in the reported temperature dependence of mercury vapor

86

pressure discussed below. To this end, we measure mercury vapor concentration in contact with

87

liquid mercury in the absence as well as in presence of air or nitrogen. We note that for a variety

88

of reasons, use of an open system containing flowing air saturated with mercury could be prone

89

to transport losses. Possible loss mechanisms include those caused by sampling, incomplete

90

saturation, thermal instability, or chemical reaction, among others. In the following study, we


Page 5 of 29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


91

have therefore used closed systems in the form of a static sealed quartz cell loaded with either

92

pure liquid mercury only or liquid mercury in presence of atmospheric pressure air or nitrogen.

93

This arrangement ensures gaseous elemental mercury stays in continuous equilibrium over a pool

94

of liquid mercury, and is enclosed in an optically transparent and chemically inert material.

95

Inclusion of liquid mercury in equilibrium with nitrogen gas near atmospheric pressure gives us a

96

measurement "sample triad". This strategy allows us to screen for the confounding effects of

97

Hg-O2 photosensitization17-19 and possible laser fluence population-saturation threshold

98

requirements needed for accurate measurement of the Hg-air system vapor concentration. The

99

resulting detection parameters help minimize, within our experimental uncertainty, the

100

occurrence of these effects by using low ultraviolet laser intensity; and is described in the

101

supplementary information. Our experiments use ultra-zero grade commercial source air

102

(synthetic high purity oxygen, nitrogen mixture) for the air buffer gas and are not subject to

103

photosensitization reactions that could occur in the presence of common flue-gas emission

104

components18. Because of the high spectral resolution of the measurements, we can account for

105

changes in the spectra caused by collisional effects. This capability enables us to precisely

106

measure relative Hg vapor concentrations in the presence or absence of surrounding air.

107

Furthermore, we can compare our measured absolute mercury vapor concentration values to

108

those predicted by thermodynamic and empirical correlations. This comparison allows us to

109

establish which of these formulas provides the most accurate representation of Hg vapor pressure

110

for Hg-air standard generators.

111

The remainder of this article is organized as follows. First, we describe the experimental

112

system and the line-by-line spectroscopic model used for data reduction. We present measured

113

and fitted spectra for the Hg-only, Hg-Air, Hg-N2 systems and discuss sources of bias. We



Page 6 of 29

114

discuss spectrometer performance metrics including detection limits, measurement uncertainties

115

and potential application to absolute mercury isotope ratio measurements. Finally, we report

116

room-temperature measurements of mercury concentration at room temperature for the Hg-only,

117

Hg-Air, Hg-N2 systems and compare results with literature values and predictions.

118

EXPERIMENT

119

We used a commercial, continuous-wave (CW) light source, consisting of a tunable, narrow

120

linewidth (sub-MHz) laser system to probe the 6 1S0 ← 6 3P1 absorption transitions of elemental

121

mercury near the UV wavelength λ = 253.73 nm. The light source is a tapered-amplifier, fourth

122

harmonic generator (TA-FHG), which comprises an external cavity diode seed laser emitting at

123

λ = 1.015 µm, followed by a semiconductor amplifier and two cascaded second-harmonic

124

generation stages for non-linear frequency quadrupling to the UV region. The resulting single-

125

frequency output can be continuously scanned over more than 40 GHz and has a short-term

126

linewidth < 500 kHz. The system produces UV laser output power of more than 20 mW with a

127

specified residual intensity noise less than 0.3 % of the mean value.

128

At each spectrum step, the frequency-doubled beam near λ = 507.5 nm is measured with a

129

wavelength meter (1 MHz frequency resolution) and the transmission of the UV laser beam

130

through the sample cell is recorded. The laser and wavelength meter are used in a closed-loop

131

configuration to actively stabilize the probe laser frequency to the desired setpoint values.

132

Spectral scans were collected over 40(80) GHz at UV frequency steps ranging from 200 MHz to

133

1 GHz (400 MHz to 2 GHz) depending on the sample system and desired resolution. The laser

134

frequencies at each setpoint had a standard deviation of approximately 2 MHz over the data

135

collection time. The narrow linewidth and precise frequency tunability enable the acquisition of

136

highly resolved absorption spectra (MHz-level precision) that were analyzed as discussed below. 6 ACS Paragon Plus Environment

Page 7 of 29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


137

The experimental configuration is schematically shown in Figure 1. The mercury vapor

138

absorption cells were custom fabricated to have a quartz cylindrical body with parallel but tilted

139

(2°) faces, thermally sealed with fused silica wedged (2°) windows. This geometry minimized

140

back reflection and etalon effects. The cell path length (nominal value of 10 mm) was measured

141

at several orientations and had an overall relative standard uncertainty of 0.1 % about its mean

142

path length (NIST, Dimensional Metrology Group). The cells contained a small volume of liquid

143

mercury localized in a short stem extending below the main cell body. The sensitivity of path

144

length to cell orientation and multiple reflections is provided in supporting information, S3.

145

Three separate sealed cells containing mercury in the absence of added gas (p < 10-6 Pa),

146

added air (ultra-zero grade) and N2 (99.9999% purity) both at 80 kPa nominal pressure were used

147

for the present studies. These cells contain liquid mercury at natural isotopic abundance in

148

equilibrium with its vapor in the surrounding gas and are subsequently referred to as Hg-only,

149

Hg-Air, and Hg-N2, respectively.

150

calibrated against an ITS-90 standard platinum resistance thermometer (NIST, Thermodynamic

151

Metrology Group) was used to monitor the cell temperature. All measurements were conducted

152

at room temperature near 22 °C.

A four-wire industrial platinum resistance thermometer

153

The laser beam was intensity modulated with an acousto-optic modulator (AOM) at a

154

modulation frequency of 20 kHz and used in first-order. The intensity-modulated laser beam

155

was expanded and then split into two paths (sample and reference legs), where each path was

156

terminated by a GaP photodiode with a bandwidth of 300 kHz under a load resistance of 1 kΩ

157

and noise-equivalent-power of 13 fW/Hz1/2. With this arrangement, the sample path was used to

158

irradiate the sample cell and the reference path was used for signal normalization to compensate

159

for variations in the laser intensity. The sample-path and reference-path photodiode signals were 7 ACS Paragon Plus Environment


Page 8 of 29

160

measured using separate phase-sensitive (lock-in) amplifiers, both of which were referenced to

161

the same AOM modulation signal. A multifunction data acquisition device was used to digitize

162

the voltage outputs. Customized software was written to record the signal and control the

163

stepwise frequency tuning of the laser across mercury absorption profile.

164

characterization studies, the total power and spatial profile of the beam incident on the cell were

165

measured using a calibrated power meter and camera beam profiler, respectively.

For saturation

166

SPECTROSCOPIC MODEL

167

6 1S0 ← 6 3P1 transition mercury isotope absorption spectra near λ = 253.7 nm

168

It is well established that the relatively high mercury absorption cross section of the 6 1S0 ←

169

6 3P1 intercombination transition provides an extremely sensitive measurement of mercury

170

concentration.

171

GEM concentration by probing this transition. Although, the spectral resolution was limited by

172

the linewidth of their pulsed laser, ∆ν = 7.5 GHz, they were able to demonstrate ultrasensitive

173

detection of GEM in ambient air at a molar fraction of 7 pmol mol-1. Subsequently, Fain et al.13

174

developed a prototype CRDS spectrometer using a pulsed laser with a linewidth of 0.9 GHz

175

which enabled them to resolve the Doppler-broadened spectra of mercury and to achieve a

176

detection limit of 0.04 pmol mol-1 for an averaging time of 1 s. This work was extended by

177

Pierce et al.15 to incorporate a frequency-locked probe laser with the CRDS method for

178

automated field-deployable measurement of GEM.

13-16,20-24

Jongma et al.14 first demonstrated the application of CRDS to measure

179

The emergence of narrow linewidth, tunable CW diode lasers combined with non-linear

180

frequency conversion schemes for the generation of UV radiation has enabled measurements of

181

the 6 1S0 ← 6 3P1 intercombination transition with improved spectral resolution well below the 1

182

GHz level. Specifically, Alnis et al.21 were the first to resolve the Doppler-broadened mercury 8 ACS Paragon Plus Environment

Page 9 of 29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


183

isotope lines at 253.7 nm. They used sum frequency generation (SFG) of visible diode laser

184

beams to generate 1 nW of UV power at 254 nm with a tuning range of 35 GHz. Carruthers et

185

al.23 significantly improved the SFG conversion efficiency achieving output powers of 50 nW

186

and demonstrated a tuning range of 70 GHz which allowed them to probe both the Doppler- and

187

air- broadened spectra of mercury with a resolution of approximately 10 MHz. Anderson et al.22

188

subsequently developed a ruggedized diode-laser-based sensor package for mercury detection at

189

253.7 nm. Their system achieved a rapid spectral scan rate (up to 4 kHz) as well as a tuning

190

range of 100 GHz to fully capture the air-broadened absorption spectra of mercury. Based on an

191

inspection of published spectra Anderson et al. demonstrated the highest spectral resolution and

192

signal-to-noise ratio (SNR) when compared to Alnis et al. and Carruthers et al. More recently

193

Lou et al.24 used a frequency-doubled multimode diode laser as a simple and rugged laser source.

194

However, for quantitative measurements, the multimode character of the probe laser required a

195

sample-reference cell configuration to extract the concentration of GEM. The potential for long-

196

term (14 h) frequency stabilization and signal averaging was recently exploited by Almog et al.20

197

using Doppler-free saturation spectroscopy of this transition.

198

As discussed below, the absorption spectrum of pure mercury vapor at room temperature and

199

natural isotopic abundance for the 6 1S0 ← 6 3P1 intercombination transition exhibits six peaks

200

originating from a total of 10 transition lines spanning approximately 22 GHz (supporting

201

information, S1). When the mercury vapor mixes with a bath gas such as air or N2 the spectrum

202

(as will be shown) the spectrum is dramatically altered due to pressure-broadening-induced

203

blending of individual component transitions.



Page 10 of 29

204

Beer-Lambert Law and Signal Analysis

205

We begin by assuming that the composite absorption spectrum can be modeled from first

206

principles using the Beer-Lambert law, which relates the absorption coefficient, α (ν ) , to the

207

product of the absorber number density, n, times its local cross section σ (ν ) , where ν is the

208

optical frequency.

209

The dual-path arrangement of Fig. 1 yields a pair of electronic signals sr and ss, which

210

correspond to the reference and signal channels, respectively. Assuming that the reference

211

photoreceiver responds linearly to the laser irradiation then sr = Piεrℜr , in which Pi is the beam

212

power at the location where the incident beam splits into the two paths, ε r is the fraction of this

213

power that reaches the reference detector, and ℜr is the transimpedance gain of the reference

214

detector and phase-sensitive amplifier system. Similarly, for the sample channel and in the

215

absence of absorption then ss = Piεsℜs . For an absorbing sample of path length l , the Beer-

216

Lambert law gives the transmittance as Ts = exp(−α(ν )l) . Dividing the sample signal by the

217

reference signal, yields the normalized transmission signal as,

z(ν ) =

ss Piε sℜs exp(−α (ν )l) = sr Piε r ℜr

β (1 + κ∆ν1)exp(−α (ν )l) ,

(1)

218

in which β and κ are measured system constants and ∆ ν is the frequency detuning relative to

219

the first point of the scan. Here, β and κ are obtained by the first-order Taylor series expansion

220

of the ratio (εsℜs / εrℜr ) = f (∆ν) such that β = f (∆ν = 0) and κ = (df / dν)∆ν=0 / f (0) . The latter quantity

221

accounts for any small difference in the frequency dependence of the collection and

222

measurement efficiencies between the two paths that leads to imperfect normalization. Physical


Page 11 of 29 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


223

mechanisms contributing to this complication include slight variations in beam steering,

224

reflection efficiencies, and etalons etc. as the laser frequency is tuned. This term can also capture

225

slowly varying wavelength-dependent losses like Rayleigh scattering and the wings of distant

226

absorption lines by species such as O2. We also note that spectra acquired with a similar-sized

227

but empty cell also were conducted to rule out the contribution of any absorption features from

228

the cell and air medium.

229

For optically thin conditions that correspond to the limit αl

Development of a high-resolution laser absorption spectroscopy

Recommend Documents