Refractometry 197

CHAPTER 7

REFRACTOMETRY

“Eigentlich weiss man nur wenn man wenig weiss, mit dem Wissen wächst der
Zweifel.”
(“We know accurately only when we know little; with knowledge doubt increases.”)
Goethe

7.1 REFRACTIVITY OF AIR

When performing length interferometry in air, it is important to correct the laser

wavelength for the refractivity of the air through which it passes. The correction factor,
the refractive index, is applied to the vacuum wavelength of the light emitted by the
laser

λ vac
λ= (7.1)
n

where λ is the wavelength in air, λvac is the wavelength in vacuum, and n is the
refractive index of air, for the ambient conditions.

The refractivity [1] of a gas, (n-1), can be expressed as the product of 2 factors: the
dispersion factor Kλ and the density factor DTP. The dispersion factor depends only on
the wavelength and the density factor is independent of the wavelength. Hence

(n − 1)TP = Kλ DTP (7.2)

7.1.1 Dispersion factor

Svensson [2] demonstrated the invariance of the dispersion factor with temperature and
Erickson [3] demonstrated its invariance with pressure. Due to this invariance, the
dispersion factor can easily be determined to high accuracy by relative measurement,
and in 1953 the dispersion formula was published by Edlén [4], based on the work of
other authors’ investigations.
198 Chapter 7

n−1
0.000284
0.000282
0.00028
0.000278
0.000276
lambda Hnm L
400 450 500 550 600 650 700

Figure 7.1 - Dispersion of standard air over the region 350 nm - 700 nm

This formula for the dispersion factor was expressed in the Sellmeier form

1
(n − 1) = ∑i Ai (7.3)
σi − σ 2
2

The σi represent wave-numbers of certain resonance frequencies, σ is the experimental

wave-number. This is the basic form of the more recent dispersion formula (see below)
that is now used. The Sellmeier formula is based on the idea that the refractivity of air
is due to discrete absorption frequencies. This can be modelled by using Maxwell’s
equations in a dielectric medium in which an electric field interacts with an idealised
harmonic oscillator, with various resonant frequencies. This is, however, not quite
correct, as these absorption bands are not discrete frequencies, but continua. The term
with the constant 38.9 in the formula below represents the effect of the absorption
continuum of O2, peaking at 145 nm, with a range of 40 nm. The constants 130 and
8342.13 represent the effect of the absorption continua of N2 and O2, starting at 100
nm, levelling off at 50 nm, before tailing off for lower wavelengths. Thus there is an
empirical equation for Kλ , the dispersion factor, based on measurements of the
dispersion of air.

7.1.2 Density factor

Edlén [5] gives a derivation for the density factor, DTP based on the Lorentz-Lorentz
equation, and the gaseous equation of state

PV
= 1 − εT P (7.4)
RT
 Refractometry 199

PV
and measured values of for dry air at different conditions. The form of the derived
RT
equation is
P(1+ β T P)
DTP = (7.5)
(1+ α T )

This corresponds to equation (7.7) below relating the refractivity of air at non-standard
pressures and temperatures, to that at standard conditions.

7.2 EDLÉN’S EQUATIONS FOR THE REFRACTIVITY OF AIR

In 1965 Edlén reviewed the most recent work [6,7], collated findings and issued new
formulae for the dispersion of air [5]. The formulae derived in that paper have since
been widely used to correct for the refractivity of air, with a minor correction to the
humidity term suggested by Birch and Downs [8].

7.2.1 Refractivity of standard air

According to Edlén, standard air (dry air at 1 atmosphere, 15°C, containing 300 ppm by
volume of CO2) has a refractivity of

2406030 15997
(n − 1)s × 108 = 8342.13 + 2 + (7.6)
130 − σ 38.9 − σ 2

where n is the refractive index, σ is the vacuum wave-number in µm-1. As an example,

for air at 15 °C, λ = 633 nm, (n-1)s = 0.000 276 517.

This is found to be accurate to about 1 x 10-8 over the wavelength range from 200 nm
to the infra red, and represents the dispersion factor, described above.

7.2.2 Corrections for temperature, pressure, water vapour and CO2

The effects of temperature and pressure are calculated from the density factor

P  1+ P(0.817 − 0.0133T ) × 10 −6 
(n − 1)TP = (n − 1)s × (7.7)
720.775  1+ 0.0036610T 
200 Chapter 7
where P is the air pressure in torr, and T is the air temperature in °C. At 1 atmosphere
(1013.25 mbar, 760 torr) and 20 °C, (n-1)TP = 0.000 271 786.

A correction can also be applied for water vapour (humidity)

( )
nTPf = nTP − f 5.722 − 0.0457σ 2 × 10 −8 (7.8)

for air containing f torr of water vapour.

Birch and Downs [8] suggest an amendment to the last equation, whereby it becomes

nTPF = nTP − f (3.7209 − 0.0343σ 2 )× 10−8 (7.9a)

but with f in mbar, not torr. When converted for measurements of f in torr, the
wavelength-dependent term is unchanged from Edlén and the equation becomes

nTPF = nTP − f (4.9608 − 0.0457σ 2 )× 10 −8 (7.9b)

This is to correct an error in the water vapour term discovered using an absolute gas
refractometer. Muijlwijk [9] suggested alterations to the standard formulae, on the
grounds that normal laboratory conditions are now typically 20 °C, and 400 ppm by
volume of CO2. However, he acknowledged the fact that the equations that he presented
agreed with Edlén’s original equations to within 1 part in 108, and thus for reasons of
standardisation with other existing instruments, the refractivity equations used in the
work of this thesis are those due to Edlén, 1966, with the Birch and Downs
modification to the humidity term. The Birch and Downs revised humidity term has
been verified by Beers and Doiron [10].

Birch and Downs expect the uncertainty of their modified Edlén equations to be
± 3 x 10-8 unless more accurate measurements of the pressure, temperature and
humidity are made. The amount of Carbon Dioxide must also not deviate from the
standard value of 300 ppm by volume, unless its value is measured directly, as is the
case with the NPLBI.

The effect of CO2 concentration on the refractivity, according to Edlén, is given by

(n − 1) x = [1+ 0.540(x − 0.0003)](n − 1)s (7.10)

where x is the CO2 concentration, by volume. The effects of other gases commonly
found in the atmosphere are negligible, unless specific contamination is introduced.
 Refractometry 201

7.3 EFFECTS OF PRESSURE, TEMPERATURE, HUMIDITY AND CO2 ON

REFRACTIVITY

Table 7.1 summarises the effects of the air temperature, pressure, water vapour content
(humidity) and CO2 content on the refractive index.

Parameter Typical value Typical daily variation in Effect of variation on

calibration laboratory refractive index

temperature 20 ° C ± 0.2 °C m 1.85 x 10-7

pressure 1013.25 mbar ± 20 mbar ± 5.36 x 10-6

humidity 13 mbar ± 2 mbar m 5.45 x 10-8

CO2 content 300 ppm ± 100 ppm ± 1.47 x 10-8

Table 7.1 - Typical variations in air temperature, pressure, humidity and CO2 content

Thus air pressure and temperature have the largest effect on the refractive index, hence
accurate measurement of these parameters is essential if the refractive index is to be
calculated accurately using Edlén’s equations.

Parameter Increment Effect on refractive index

temperature 1°C -9.26 x 10-7

pressure 1 mbar 2.68 x 10-7

humidity 1 mbar -2.73 x 10-8

CO2 content 100 ppm 1.47 x 10-8

Table 7.2 - Effect of air temperature, pressure, humidity and CO2 content on refractivity

7.3.1 Pressure measurement

The pressure transducer used in the PLBI is a Druck DPI 140. This instrument has a range
of 35 - 1150 mbar (absolute), with a resolution of 0.01 mbar. The manufacturer’s
estimated accuracy of the instrument is ± 0.02% of reading ± 0.01% of full scale, or
approximately ± 0.23 mbar, though with direct calibration against NPL primary
standards, an estimated uncertainty of ± 0.08 mbar can be achieved.

The DPI 140 is situated in the equipment rack and is connected via a tube to the
chamber. The opening of the tube is positioned on the carriage, close to the length bars,
202 Chapter 7
ensuring that the pressure sense point is close to the measurement beam. To prevent the
pump used to sample the chamber air from affecting the pressure, it is switched off
before the pressure readings are made. The DPI reading is available over the IEEE bus,
and is read by the computer, which applies a software correction to the reading
according to the latest calibration.

The DPI works by measuring the resonant frequency of an oscillating cylinder, which is
dependent on the density of the air surrounding the cylinder, which in turn is dependent
on the air pressure. The measurement is thus a measurement of density and is affected,
to second order, by humidity and other gaseous contaminants. The effect of these has
been minimised by the used of the sealed chamber, situated in a humidity-controlled
laboratory. From the manufacturer’s handbook, it is seen that the errors given in table
7.3 arise in the pressure reading due to a change from 0% Relative Humidity (RH) to
70%, at the stated temperatures.

Temperature (°C) Error (% reading) Error at 1 atm (mbar)

10 0.004 0.04

25 0.010 0.10

30 0.013 0.13

Table 7.3 - Errors in pressure reading due to change from 0% to 70%RH

7.3.2 Temperature measurement

The temperature of the air inside the chamber is measured in situ by a platinum
resistance thermometer. The PRT is enclosed in a small heatsink, placed on an insulating
support, close to the length bars. The PRT thus measures the temperature of the air
within a small region, which requires the temperature along the measurement beam to
be uniform. An alternative would be to use many PRT sensors along the path, but the
extra time taken to measure these sensors (approximately 20 seconds each) would allow
drift of the readings of temperature and pressure and thus outweigh the advantage of
having more measurement points. The temperature gradients along the beam paths are
very small and do not affect the air temperature measurement: when the chamber is
temperature-controlled at 20 °C, the temperature gradient along the measurement path
has been measured as less than 0.001 °C per metre, and hence only one temperature
measurement is required. The drift rate is less than 0.002 °C per hour.

The resistance of the PRT is measured using a Tinsley Senator 5840D precision AC
resistance bridge which is based on an NPL design. The bridge uses a series of
computer-switched precision transformer windings to balance the voltage in the two
 Refractometry 203
arms of a Wheatstone bridge. The decade-windings are switched in or out according to
the voltage difference between the two arms. The 1:1 windings are used first, followed
by the 10:1, then the 100:1 and so on, gradually decreasing the out-of-balance signal. A
digital display can be configured to show either the resistance reading, resistance ratio
to an external standard resistor, or temperature calculated by an external computer,
connected via the IEEE bus, along which the resistance reading or an out-of-balance
indicator signal can be sent.

The PRT resistance is measured in a 4-terminal configuration to account for the

resistance of the cables. The bridge contains a standard resistor, which is temperature
controlled. The resolution of the bridge is 0.000 01 ohms, corresponding to a
temperature resolution of 0.000 03 °C. When calibrated against an external 100 ohm
standard resistor, the accuracy of the bridge is estimated to be ± 0.000 47 °C (0.47 mK).

Each PRT used with the bridge is individually calibrated at the water triple point and the
melting point of gallium, 0.01 °C and 29.7646 °C, respectively. The uncertainties in the
realisations of these standard temperatures are ± 0.000 5 °C and ± 0.000 5 °C,
respectively. Calibration of the resistance of the PRTs at these temperatures allows
calculation of corresponding resistance-temperature coefficients according to the
International Temperature Scale of 1990 (ITS-90). Following a measurement of the
resistance of a PRT, the ITS-90 coefficients allow the corresponding temperature to be
calculated to within an accuracy of 0.000 13 °C (see § 8.3).

7.3.3 Humidity measurement

The water vapour content (humidity) of the air is measured by a Michell S3000
dewpoint hygrometer. This device measures the dewpoint of air passing over a
temperature-controlled mirror by controlling the temperature of the mirror to the point
where dew forms on its surface. A photocell senses the intensity of light reflected from
the mirror surface - this signal decreases when dew forms on the mirror. The
temperature of the mirror is monitored by a PRT. A simple formula [11] is used to
convert the dewpoint temperature, Tdp, into a partial pressure of water vapour, f, in
mbar.

 7.5 × Tdp 
f = log  + 0.78571
−1
(7.10)
 237.3 + Tdp 
204 Chapter 7
The resolution of the S3000 is 0.1 °C dewpoint and is calibrated against traceable
humidity standards to provide an accuracy of ± 0.25 °C. The dewpoint reading is
converted into a voltage which is read by a DVM with an IEEE interface. Overall
accuracy of humidity measurement is estimated to be ± 0.3 °C (± 0.24 mbar vapour
pressure).

Due to the heat output of the peltier cooler attached to the dewpoint mirror, the sensor
is mounted outside the chamber. Air is sampled from the chamber from a point near the
length bars by a PTFE-coated re-circulating pump. The air flows along PTFE/stainless
steel tubes, through the S3000 sensor, through the CO2 sensor, and is returned to the
chamber. The pump surfaces are PTFE and are sufficiently clean for analytical gas
sampling operations and hence do not contaminate the returned air. The flow rate of the
sampled air is approximately 0.5 litres min-1.

7.3.4 Carbon dioxide measurement

The CO2 content of the sampled air is measured using an Edinburgh Instruments
GASCARD CO2 sensor. This measures the relative absorption of light passing through a
sample tube of gas. The sampled gas flows through a single chamber, approximately
10 cm long, though which a wide-band light source shines. The intensity detected at
certain wavelengths is used to provide an output signal (monitored by an IEEE DVM)
which corresponds to the CO2 molar density. The instrument has a resolution of
± 18 ppm CO2 and an uncalibrated accuracy of ± 60 ppm. The device is calibrated
using standard gases at 0 ppm and 370 ppm (by volume) to achieve an estimated
accuracy of ± 30 ppm.

7.3.5 Air parameter measurement order

The measurements of air temperature, pressure, humidity and CO2 concentration are
performed at the same time as the phase-stepping measurements. The order of the
individual measurements is given below. The order of measurement of devices is
designed so that parameters which are slowly varying (humidity, CO2, bar temperature)
are read at the start and end of the measurement, whereas more quickly varying
parameters (air pressure and temperature) are read at the same time as the phase-
stepping is performed. The CO2 and humidity readings are taken before the sample
pump is temporarily halted.
 Refractometry 205

(1) The air temperature PRT is selected.

(2) The humidity sensor is read by the computer.
(3) The CO2 sensor is read by the computer.
(4) The sample pump is stopped.
(5) The air temperature PRT is read.
(6) One of the two PRTs for measuring the temperature of the length bar is read.
(7) The phase-stepping is performed at three wavelengths.
(8) The air pressure is read by the computer.
(9) The second bar temperature PRT is read.
(10) The sample pump is re-started.

From the measurements of pressure, temperature, humidity and CO2 content, the
refractivities at the three measurement wavelengths are calculated using the modified
Edlén equations. The wavelengths corrected for the effects of refractive index are used
in the length calculation routine.

7.4 OTHER WAVELENGTH COMPENSATION TECHNIQUES

7.4.1 Two-wavelength compensation

Matsumoto and Honda [12] used a two-wavelength interferometer based around a YAG
laser with a frequency-doubling crystal (KTiOPO4) producing both 1.06 µm and 532
nm wavelengths. The interferometer averaged fringe counts at the two wavelengths to
derive a wavelength-corrected fringe count which corresponded to the length being
measured by the moving part of the interferometer. An accuracy of length measurement
of 0.2 µm over 1 m of travel was obtained in air.

7.4.2 Gas refractometry

The refractive index of air can be measured directly in a gas refractometer. The
refractive index, or the refractivity (the refractive index -1) of a sample of gas is
measured relative to that of a known standard, often vacuum.

In order to check the accuracy of the results of the modified Edlén equation and as a
possible second technique for refractive index determination, a gas refractometer was
206 Chapter 7
constructed at NPL by Mr P D West, based on the design of the NPL Gas Refractometer
of Birch and Downs [13,8].

7.4.2.1 Gas refractometer design

The design of the refractometer is shown in figure 7.2.

Mirror
Polarised laser

Photo
A detectors
C

End End
window Cell window
λ/4

Jamin Corner
beamsplitter cube
in out out in

Figure 7.2 - Gas refractometer schematic

This instrument is based around a polarised interferometer using a Jamin beamsplitter

and a linearly polarised laser. A concentric two-chamber stainless steel gas cell with
quartz end plates is used to contain the sample and reference gases. The inner cell is
used for the reference, which is usually vacuum and the outer cell for the sample gas.
The external solenoid valves and pipework allow interconnection of the two chambers
allowing serial flow of the sample gas though both chambers, gas-outer & vacuum-
inner, or both vacuum.

Vacuum is provided by a simple rotary pump (Edwards Model 1.5), which achieves a
vacuum of 3 x 10-2 mbar. The three photo detectors are gain matched: two provide the
orthogonal output signals from the interferometer, the third is used to adjust the gain of
the other two detection channels, based on the intensity of the laser beam.

The internal length of the cell is approximately 316.6 mm, corresponding to 106
interference fringes, at the laser wavelength, thus when comparing the optical path
lengths of the two cell chambers, a difference in refractive index of 1 x 10-6 will be
equivalent to 1 interference fringe. This is the basis of operation of the interferometer.
 Refractometry 207

7.4.2.2 Refractometer operation

Sample gas is flowed though both chambers using the sampling pump for a period of
10 minutes. The pump is then switched off to allow the gas to stabilise at ambient
pressure. A fringe counter, connected to two photocells which monitor the two
orthogonal outputs of the interferometer, is zeroed. The sample pump is re-started and
the inner chamber evacuated. As the air is removed, the change in refractive index
causes a change in optical path length of the reference cell. The signals from two
orthogonal outputs then exhibit a sinusoidal variation corresponding to the passage of
interference fringes across the two detectors. The fringe counter tracks these fringes by
tracking the two outputs.

When a sufficient vacuum has been obtained in the inner chamber, the display of the
fringe counter will be the refractivity of the sample gas, with a factor of 10-6. When a
reading is required, the sample pump is stopped for 60 secs before the reading is taken.
As an example, a fringe count of 267.123 corresponds to a refractive index of
1.000 267 123.

7.4.2.3 Corrections and errors

Due to constructional and operational constraints, a number of corrections have to be

applied to the reading of the fringe counter to obtain the correct refractive index of the
sample gas.

(1) Cell length

Due to engineering tolerances, the cell length is not exactly 1 million half-wavelengths
long, thus the fringe count is first multiplied by a correction factor of 0.999716
(determined from accurate measurements of the cell length using a Co-ordinate
Measuring Machine and a Gauge Block Comparator).

(2) Imperfect vacuum

The refractive index of absolute vacuum is 1 by definition. The refractive index

increases by 2.7 x 10-9 per 10-2 mbar, thus due to the imperfect vacuum achieved by the
rotary pump, a correction of 8.1 x 10-9 is added to the fringe count.
208 Chapter 7

(3) Cell expansion

The cell is made of stainless steel which has a thermal expansion coefficient of
14.7 x 10-6 K-1. Thus the cell length is dependent on temperature. A correction factor of
5 x 10-9 K-1 departure from 20 °C is applied to the fringe count.

(4) Gas adsorption

It has been shown [14] that when moist air is admitted to a previously evacuated cell in
order to measure its refractivity, some water molecules are absorbed by the metallic
walls of the cell, leading to a decrease in the humidity of the air with a corresponding
increase in the refractivity of up to 1 x 10-7. This effect can be removed by
continuously flushing with ten cell volumes of gas, prior to making a measurement.
Alternatively, the cell can be used in the reverse configuration with the two chambers
of the cell initially flushed with air before one is evacuated (as used in the refractometer
used with the NPLBI).

However, the process of physisorption [15] can still lead to an error. Here, water
molecules in the stationary air are adsorbed onto the stainless steel wall, leading to an
increase in the refractivity of up to 1.5 x 10-7. The effect can be minimised by
increasing the flow rate of the gas prior to stopping the flow for measurement and by
reducing the overall size of the refractometer cell tubes.

(5) Pressure effects

Unless the pressures at the sample point and within the cell are equal, there will be a
pressure-dependent refractivity gradient between these two points, and the refractivity
measured by the refractometer will not be the same as that at the sample point. Thus,
the effect of gravity on the air pressure must be taken into account, and the cell placed
at the same height (above sea level) as the sample point. The earth’s pressure gradient
near the surface causes a change in refractive index of -3 x 10-8 m-1 ascent.

A second order pressure effect is due to the optical path length changes induced in the
cell windows by the pressure change as air is evacuated from one of the cell chambers.
For a 320 mm long cell, this has been shown [16] to cause a change in the measured
refractivity of up to 1 x 10-8.
 Refractometry 209

(6) Temperature effects

The refractometer measures the refractive index of the gas under the conditions present
inside the cell chamber. If there is a temperature gradient between the sample point and
the cell, then the refractive index will be different at these two points and a correction
factor, based on Edlén’s equation will be required. When the length bar interferometer
is being used to perform thermal expansion measurements, the temperature of the air
inside the sealed chamber can be as high as 30 °C. The temperature of the air
surrounding the refractometer is approximately 20 °C. The specific heat capacity of air
is so low, that the air sampled from the interferometer chamber cannot supply enough
energy to heat the refractometer cell by more than approximately 0.5 °C, and so a large
temperature gradient exists between the chamber and the refractometer cell. This will
cause a change in refractive index due to not only the temperature change (which can
be measured) but also a possible change due to possible outgassing of the connection
pipes or water condensation inside the cell, which cannot be measured.

(7) Fringe interpolation

As mentioned in section 2.3.1, errors arise in the fringe interpolation of the fringe
counter due to effects such as gain mis-match between the two signals, offsets and
phase errors. Some degree of computer correction is possible [17] by fitting ellipses to
the data, although this takes time and requires low-noise signals. In the present
refractometer, drift of the electronics or the alignment of the interferometer causes the
offset and gain for the two channels to vary, leading to fringe interpolation errors.

7.4.2.4 Comparison between Edlén and refractometer

To compare the performances of the modified Edlén equations and the refractometer,
they were both used to measure the refractive index of air sampled from the
interferometer chamber.
210 Chapter 7

PRTs
FLOW
Interferometer
chamber Refractometer

Humidity sensor
Pressure sensor

Pump CO2 sensor

RETURN

Figure 7.3 - Comparison of Edlén and refractometer

The value of the refractive index measured by the refractometer was compared against
that calculated using the modified Edlén equation based on measurements of the air
parameters. The air temperature used for the calculations was that inside the
refractometer cell and the pressure was measured inside the interferometer chamber.

General agreement between the two methods was ± 3 x 10-8, as can be seen in figure
7.4. On analysis of the data, it can be seen that the difference between the two methods
is periodic with a period equal to one refractometer fringe, as shown in figure 7.5. This
indicates that the two channels do not remain in perfect phase quadrature with equal
offsets and gains. This was also detected by viewing the outputs of the two channels as
a lissajous figure on an oscilloscope.

Figure 7.4 - Comparison of Edlén and refractometer

 Refractometry 211

2e-7

1e-7
Refractometer-Edlén

0e+0

-1e-7

92.11.20
-2e-7
1.000265 1.000266 1.000267 1.000268 1.000269 1.000270 1.000271
Refractometer reading

Figure 7.5 - Periodic error due to drift

Thus although an absolute gas refractometer can be used to measure the refractive index
of the air, the accuracy of measurement is generally no better than that achievable by
the use of Edlén’s equations coupled with sensitive, accurate transducers. Because of
the fact that the composition of the air inside the chamber does not change rapidly, if at
all, the calculation of the refractive index should be at least as accurate as the
determination using the refractometer. At elevated temperatures, unless the system is
re-designed with an internal refractometer, of a more stable design, Edlén is the only
accurate solution to measure the refractivity of the air inside the interferometer
chamber.

Birch et al [18] compared the results of air refractivity measurements made using an
automatic refractometer with those obtained using commercial sensors and the Edlén
equations. Their conclusion was that the modified Edlén equation is the most accurate
form to use for calculating air refractivity. An uncertainty of ± 3 x 10-8 is achievable.
When account is taken of excess CO2 levels (above the nominal 300 ppm assumed by
Edlén), the modified Edlén equation is accurate to about ± 1 x 10-8.
212 Chapter 7

7.5 ADDENDUM

Birch & Downs [19] have recently published updated Edlén equations for the refractive
index of air. These new formulae take into account: the adoption of SI units (the pascal
is now the recommended unit of pressure, not the torr); the establishment of the
International Temperature Scale of 1990, superseding IPTS-68; improvements in the
equation for the density of moist air [20]; increased CO2 levels; and revision of the
water vapour term. The new standard equations are as follows.

P( n − 1)s 1 + 10 −8 (0.601 − 0.00972T )P

(n − 1)TP = × (7.11)
96095.43 1 + 0.0036610T

2406294 15999
(n − 1)s × 10 −8 = 8343.05 + −2 + (7.12)
130 − σ 38.9 − σ −2

nTPf = nTP − f (3.7345 − 0.0401σ −2 ) × 10 −10 (7.13)

where T is the temperature in °C, f is the water vapour pressure in Pa, σ is the
wavelength in µm and P is the air pressure in Pa. These new formulae are expected to
have an uncertainty, over the wavelength range 350 - 650 nm, of ± 1 x 10-8. These
equations will be incorporated into the software which calculates refractive index inside
the interferometer.

[2002 re-publication note

Equation 7.12 has since been corrected by Birch & Downs to become:

2406147 15998
(n − 1)s × 10−8 = 8342.54 + +
130 − σ −2
38.9 − σ − 2

There have been several later revisions of these equations by other authors]
 Refractometry 213

REFERENCES FOR CHAPTER 7

[1] Barrel H & Sears J E The refraction and dispersion of air for the visible
spectrum Phil. Trans. Roy. Soc. London A 238 (1939) 1-64
[2] Svensson K Arkiv Fysik 16 (1960) 361
[3] Erickson K E Investigation of the invariance of atmospheric dispersion with a
long-path refractometer J. Opt. Soc. Am. 52 (1962) 777-780
[4] Edlén B The dispersion of standard air J. Opt. Soc. Am. 43 (1953) 339-344
[5] Edlén B The refractivity of air Metrologia 2 (1966) 71-80
[6] Rank D H, Saksena G D & McCubbin T K Measurements of the dispersion of
air from 3651 to 15300 Angstroms J. Opt. Soc. Am. 48 (1958) 455-458
[7] Peck E R & Khanna B N Dispersion of air in the near-infrared J. Opt. Soc. Am.
52 (1962) 416-419
[8] Birch K P & Downs M J The results of a comparison between calculated and
measured values of the refractive index of air J. Phys. E: Sci. Instrum. 21 (1988)
694-695
[9] Muijlwijk R Update of the Edlén formulae for the refractive index of air
Metrologia 25 (1988) 189
[10] Beers J & Doiron T Verification of revised water vapour correction to the
refraction of air Metrologia 29 (1992) 315-316
[11] British Standard BS 1339 (1965) (London: British Standards Institution)
[12] Matsumoto H & Honda T High accuracy length-measuring interferometer using
the two-colour method of compensating for the refractive index of air Meas. Sci.
Technol. 3 (1992) 1084-1086
[13] Downs M J & Birch K P Bi-directional fringe counting interference
refractometer Precis. Eng. 5 (1983) 105-110
[14] Birch K P & Downs M J Errors in the determination of the refractive index of air
Appl. Opt. 28 (1989) 825-826
[15] Birch K P, Downs M J & Ward R E The measurement of humidity variations in
gases resulting from the dispersion of water onto surfaces J. Phys. E: Sci.
Instrum. 21 (1988) 692-694
[16] Birch K P, Downs M J & Ferris D H Optical path length changes induced in cell
windows and solid etalons by evacuation J. Phys. E: Sci. Instrum. 21 (1988) 690-
692
214 Chapter 7

[17] Birch K P Optical fringe subdivision with nanometric accuracy Precis. Eng. 12
(1990) 195-199
[18] Birch K P, Reinboth F, Ward R E & Wilkening G The effect of variations in the
refractive index of industrial air upon the uncertainty of precision length
measurement Metrologia 30 (1993) 7-14
[19] Birch K P & Downs M J An updated Edlen equation for the refractive index of
air Metrologia 30 (1993) 155-162
[20] Giacomo P Equation for the determination of the density of moist air (1981)
Metrologia 18 (1982) 33-40