 |
Description  |
|
|
FIELD OF THE INVENTION
This invention relates to determining the composition and concentration of
an arbitrary mixture of gases in a patient's airway, and more particularly
to such determination by means of Raman spectroscopy.
BACKGROUND OF THE INVENTION
Measurement of the composition of a gas mixture is especially important in
the airway of an individual who is connected to an auxilliary breathing
apparatus. Examples include ventilators which assist respiration in the
intensive care unit and coronary care unit in hospitals, and anesthesia
machines in hospital operating rooms ("ORs"). These patients are often
monitored closely for their vital signs, including respiratory gas
exchange.
In the operating room, the anesthesiologist supports the patient's
respiration and also controls the patient's depth of anesthesia using
special gaseous agents. A patient airway will contain the normal
respiratory gases, such as oxygen (O.sub.2), carbon dioxide (CO.sub.2),
and water vapor (H.sub.2 O), or a mixture of nitrous oxide (N.sub.2 O) and
one or more halogenated anesthetic agents ("HA", usually halothane,
enflurane, isoflurane, desflurane or sevoflurane). Occasionally, nitrogen
(N.sub.2) from room air infiltrates the system. Various metabolic products
and special gases, for example, to measure pulmonary function, may also be
present. Also, interferences from extraneously introduced gases such as
ethanol and isopropanol may be present. Because severe injury can result
from use of an improper gas mixture, anesthesiologists prefer to measure
the composition of the patient breathing mixture. Every component of the
mixture is important. Measurement of oxygen concentration helps prevent
hypoxia. Presence of the CO.sub.2 waveform indicates healthy gas exchange.
Measurement of the type and concentration of the various anesthetic agents
helps control and adjust the proper depth of anesthesia. The presence of
other gases can indicate leaks and possible system malfunctions.
Several Anesthesia Monitors ("AMs") exist today that perform this function.
Four different technologies compete for most of the market today.
(1) Infrared Absorption plus Oxygen. One class of AMs uses infrared
absorption to measure the halogenated anesthetics, CO.sub.2, and N.sub.2
O. While this technique is widely used, it has the disadvantages that it
is difficult to distinguish the HAs from each other, because they have
similar IR spectra that must be measured in the region .lambda.=3-4 .mu.m.
Identification of the individual agents requires measurement in the far
infrared region, .lambda.>10 .mu.m, which is more difficult. A separate
measurement, using a paramagnetic or polarographic sensor, is needed to
measure O.sub.2, which has no infrared ("IR ") spectrum. Also, separate
cells are needed to measure the HAs and the CO.sub.2 +N.sub.2 O
concentrations. In addition, these devices have no means to detect other
gases which may be present. Such devices may introduce error in the
measurements, and the devices cannot be adapted easily to measure new
agents.
(2) Mass Spectrometry. Mass spectrometers ("MS") can provide extremely
accurate measurements of gas concentrations. Historically, MS devices are
expensive and complicated instruments that require frequent calibration
and maintenance. These devices require use of a delicate vacuum system and
ion source. Typically, many operating rooms share a single MS. This
reduces the response time between measurements, and requires relatively
long sampling lines, which can distort the gas samples. In addition, the
mass spectra of the various OR gases are not unique. Nitrous oxide and
carbon dioxide have the same mass number, and isoflurane and enflurane are
isomers. Therefore, one must observe fragmentation of these molecules in
the system and employ special algorithms to distinguish the molecules. The
systems must also be protected from some other gases, including helium and
water vapor. Shared systems cannot provide continuous, breath-to-breath
analysis of airway concentrations, which is desirable. At least one
manufacturer has introduced a small, stand-alone MS, offering each OR
suite a dedicated measuring device. Although this device eliminates the
problem of multiplexing measurements for different users, and has a more
convenient design, the device still requires protection from contaminants,
and can only be configured to measure a few specific gases at a time. The
problems of high cost and distinguishing between isomers and between
identical mass spectra remain.
(3) Photo-acoustic Spectroscopy. These devices also utilize the properties
of infrared absorption to characterize gas mixtures. A precision
microphone detects pressure waves which are produced when the gas sample
absorbs IR energy. The sound level indicates the concentration. Like the
IR systems described above, these devices need a separate system to
measure oxygen concentration, and the technique cannot easily distinguish
different HAs from each other.
(4) Raman Spectroscopy. Scattering of light by the Raman effect has
received much attention from scientists since its original exposition by
C. V. Raman in 1928. Simply stated, when monochromatic light illuminates a
vibrating molecule, light scatters in a process which decreases or
increases the frequency of the scattered light by exactly the vibrational
frequency of the molecule. The shift in frequency of radiation is
characteristic of the scattering medium, and is independent of the
frequency of the illuminating radiation. Thus, measurement of the
Raman-scattered light can be used to infer the properties of the medium,
such as the chemical composition and concentration. For measurements of OR
and airway gases, this technique has the advantage that each OR gas,
including oxygen and any poly-atomic molecule, has a unique Raman
spectrum. Additionally, the Raman spectrum for a gas is usually contained
in a relatively narrow wavelength band, which simplifies detection. Thus,
Raman spectroscopy offers the promise of simultaneous measurement of all
airway gases with a single measurement and less complex technology.
In Raman scattering, a small fraction of collisions of photons with an atom
or molecule are inelastic, with a photon either giving up a small portion
of its initial energy E.sub.0 to the collision partner and scattering as a
photon of reduced energy E <E.sub.0 (Stokes waves) or the collision
partner giving up a small portion of its initial energy so that the photon
scatters with increased energy E>E.sub.0 (anti-Stokes waves). In Rayleigh
scattering of a photon with an atom or molecule, by contrast, the energy
of the scattered photon is equal to the initial energy of the photon. This
does not include light that is absorbed and re-emitted by processes such
as phosphorescence or fluorescence. In a typical scattering situation, the
ratio of intensity of Rayleigh to initial light intensity for gases might
be about 10.sup.-9 and the ratio of intensity of Raman scattered light to
initial light intensity might be about 10.sup.-12. The change in
wavelength for a Raman scattered photon of initial wavelength
.lambda..sub.0 =c/f.sub.0 and scattered light wavelength .lambda..sub.R
=c/f.sub.R is given by
.DELTA..lambda.=.lambda..sub.R -.lambda..sub.0 =(c/f.sub.R)-(c/f.sub.0),(1)
where .DELTA..lambda.>0(.DELTA..lambda.<0) corresponds to appearance of
Stokes waves (anti-Stokes waves). Substantially all scattered light at
moderate initial energies arises from Rayleigh scattering or Raman
scattering. For molecules of moderate or higher symmetry, not all modes of
molecular vibration result in Raman lines. Some of the modes of molecular
vibration produce infrared absorption lines but not Raman scattering
lines, some modes produce Raman lines but not infrared lines, some modes
produce both Raman and infrared lines, and some vibration modes produce
neither Raman nor infrared lines.
The Raman scattering cross-sections are extremely small, and the intensity
of the scattered light is very weak, as noted above, especially in gases
where the molecular number density is also relatively small (compared to
liquids and solids), and are therefore difficult to measure with accuracy.
The differential intensity of the Raman-scattered light scattered into a
differential solid angle d.OMEGA. along a differential path length dz in a
single component gas is given by the formula
dP.sub.Raman =P.sub.o n.sub.o (d.sigma./d.OMEGA.)d.OMEGA.dz,(2)
where P.sub.o is the intensity of the incident light, n.sub.o is the number
density of the scattering molecules, and (d.sigma./d.OMEGA.) is the
differential scattering cross-section in a given direction. The direction
of the scattering is also dependent on the polarization of the incident
light. If the gas contains more than one component and the components do
not interact appreciably with one another, the intensity of each Raman
line of a gas component is proportional to the concentration of that
component so that Equation (2) above can be used with a small modification
to take account of the presence of the other components.
For a given intensity of the incident radiation and sample concentration,
one can maximize the magnitude of the measured Raman signal only by
increasing the solid angle of the light-collecting optics, or by
increasing the observation path length (i.e., using a larger scattering
volume), because the molecular properties of the sample are not variable.
In a conventional system for observing Raman-scattered light, a laser beam
is brought to a focus in the medium of interest, creating a minute region
of relatively intense electrical field, which excites the Raman effect.
The light scattered from this region is collected by an optical system,
typically a simple lens which images the scattering region onto a suitable
optical filter and detector. The difficulty of improving this simple
design is evident from the observation that the etendue (defined as the
product of the collection area and collection solid angle) is conserved in
any ideal optical system. Thus, increasing the solid angle of light
collection (lower f number optics) decreases the observable area (and thus
the path length). Each experimental system must optimize these parameters
against its own constraints. Some Raman systems further improve signal
intensity by providing multiple passes of the incident light beam through
the observation volume, effectively increasing the path length, or by
placing the observation volume inside an optically resonant cavity,
effectively increasing the incident intensity P.sub.o.
Analysis of Raman scattered light is especially useful where the Raman
spectrum of each of the components present in a sample is a relatively
simple line pattern and the Raman lines of the different components do not
coincide with or lie close to one another. A complete range of vibrational
frequencies can be covered with one monitoring instrument, and the sample
container can be glass or many other relatively transparent materials.
Water may be present; the Raman spectrum of water is weak and diffuse in
the band 200-3300 cm.sup.-1, but the spectrum has a strong, broad peak
centered at 3652 cm.sup.-1. The approximately linear relationship between
component concentration and Raman-scattered light intensity makes the
calculation of concentration straightforward. Integration across a portion
of the Raman spectrum to determine the intensity of specified lines is
also straightforward. However, the sample should be non-fluorescent, and
the sample to be analyzed should be relatively transparent, with little or
no absorption at the wavelengths of interest, and should be free of
particulates. It is often difficult to apply Raman scattering to very low
concentrations of the sample, because of the weak intensity of the
individual Raman lines.
Raman systems that measure the composition in patient airways must measure
multiple spectral lines in order to distinguish all the component gases.
Several earlier patents describe techniques for measuring multiple gases.
Albrecht, in German Pat. DE 2723939C2, describes a system of six detectors
mounted in an equatorial plane around a region of focus. The focal region
has a confocal cavity where the exciting laser beam makes multiple passes
through the sample. This system requires six different detection channels,
each with associated collection optics, laser line rejection and Raman
line filters, and separate detector. The configuration is limited to six
channels and uses an unpolarized laser beam. More channels would be needed
to measure all the OR gases of interest. In addition, the confocal cavity
and each of the collection optics requires delicate adjustment to assure
proper imaging of the focal region onto the detectors.
In U.S. Pat. No. 3,704,951, Chupp discloses use of a multi-pass gas cell
for increasing the intensity level of light that is Raman scattered from a
gas sample contained in the cell. Raman scattered light exits from the
cell through a large side window in the cell.
Leonard, in U.S. Pat. No. 3,723,007, discloses a gas cell with a
transparent side window for Raman scattering analysis and notes that two
or more simple molecular gases may have distinct Raman shift spectra.
A spectrophotometer that compares light Ramnan scattered from a known gas
sample with light Raman scattered from an unknown gas sample is disclosed
by Tans et al in U.S. Pat. No. 4,630,923. Raman scattered light from the
known and unknown gas samples is alternatingly received by a detector to
determine the concentration ratios of two gases present in the unknown gas
sample.
A gas monitoring system, disclosed by Benner et al in U.S. Pat. No.
4,648,714, collects light from a single focal region, illuminated by a
properly oriented, polarized laser beam. A single set of collection optics
images the scattered light onto a single detector. An additional mirror,
placed opposite the collection optics, reflects Raman-scattered light back
into the focal region and into the collection optics, further enhancing
the signal. The sample cell can be inside the resonant cavity of the
exciting laser to increase the signal still more. The system employs a
rotating filter wheel that passes different Raman line filters in front of
the detector so that different Raman lines are measured sequentially.
While this system can measure a larger number of Raman spectral lines (by
adding more filters), the samples are not acquired simultaneously from the
same gas sample. This reduces the ability of the system to respond
rapidly, as is desired for breath-by-breath analysis of the airway
composition.
In U.S. Pat. No. 4,676,639, Van Wagenen discloses use of a gas cell for
Raman scattering analysis with transparent end and side windows. The side
windows may be coated with a narrow band anti-reflection coating, for
passage of the Raman scattered light for detection outside the gas cell.
Van Wagenen et al describe another system in U.S. Pat. No. 4,784,486, which
uses multiple detection units, each including collection optics (lens and
back reflector), laser line rejection and Raman line filters, and a
detector. Each detection unit collects Raman-scattered light
simultaneously from separate focal regions. The system is similar to the
system of Bennet, but employs a completely separate detection unit for
each Raman line, rather than inserting different filters serially. This
design has the advantage that all the channels can acquire measurements
simultaneously. However, the measurements are generated from different
sample volumes and, thus, from different gas molecules. The sample gas
must flow from location to location, and the flow rate must be such that
all units see a mixture of substantially similar composition. Thus, for a
given flow rate and size of sample cell, the practical number of detector
units is limited by the desired time response of each system. This creates
limitations because it is desirable to sample as small a volume as
possible from the patient airway. An increase in the number of detection
units increases the expense of the unit proportionally. Each unit also
requires separate alignment and adjustment, which increases the complexity
and cost of the system.
Many of the systems discussed above employ a combination of individual
narrow-band filters. Each individual filter examines only a single
spectral line or a small spectral region. This requires that a system use
at least one filter for every molecule spectral peak of interest. Allowing
for the expense of multiple channels and filters, these systems are suited
to the measurement of gas mixtures, where the Raman spectrum of each gas
consists of a single peak or a few well-separated peaks, and where these
spectral peaks are unique for each gas in the mixture, with no overlapping
spectral lines. In this case, each filter can measure each gas separately
and completely. Overlapping spectral lines can be ignored if the gas of
interest has another unique peak. This is true for the gases N.sub.2,
O.sub.2, CO.sub.2, and N.sub.2 O. The principal spectral lines of these
gases are listed below as frequency shifts from the frequency of the
illuminating source, such as a laser beam. Note that CO.sub.2 and N.sub.2
O have an overlapping line at 1285 cm.sup.-1.
N.sub.2, nitrogen: 2331 cm.sup.-1
O.sub.2, oxygen: 1555 cm.sup.-1
CO.sub.2, carbon dioxide: 1285 and 1388 cm.sup.-1
N.sub.2 O, nitrous oxide: 1285 and 2224 cm.sup.-1
The limitations of these types of systems, which employ separate filters,
arise when the Raman spectra of one molecule of interest becomes
complicated, when it covers a larger spectral band or has many, scattered
peaks, such as the halogenated anesthetics. These spectra are best
characterized by their entire spectrum, not by single lines. The Raman
spectra from the HAs have considerable spectral overlap with each other,
and the broadband emission can contribute error to the signals of the
respiratory gases and N.sub.2 O. A representative mixture of the
respiratory and anesthetic gases produces a complicated signal, with
emissions and spectral lines across the entire spectrum, a mixture of all
the individual spectra. Measurement of single spectral lines therefore
gathers less signal and less information than a technique that gathers the
entire spectral signature of the mixture or of a single component. Single
filter systems use only small amounts of the total signal available,
because such systems measure only part of the Raman emissions. For the
HAs, measurement of a single line neglects most of the total signal.
Single filter systems also neglect significant information about the Raman
spectra. It is much easier to distinguish each HA from examination of its
entire spectrum than from a single, possibly weak, line.
A further limitation of the single filter system is that new or additional
gases cannot be measured without the addition of new detection units with
new filters. New gases which have substantial spectral overlap with
existing gases may be very difficult to measure, even with additional
filters.
Also, if a Raman spectrum has closely spaced lines that must be
distinguished from each other, the spectral bandwidth of the individual
filters must be sufficiently narrow to differentiate nearby lines, or
obscure undesired lines. This places a constraint on the spectral
bandwidth of the laser source, which must be at least as narrow as the
Raman line filters, and the laser must not drift in wavelength. This
constraint also reduces the detected power. Further, filters can drift as
they age. If a filter has a bandwidth of 1 nm, and the laser drifts in
wavelength from 800 nm to 803 nm, the Raman line of interest moves out of
the band of the filter and is not measurable. For these reasons, existing
filter-based Raman systems often use gas lasers, such as argon-ion, which
has narrow, constant lines and requires high input power. Solid-state
lasers have wider line widths (>1 nm), and often require special
wavelength stabilization. Semiconductor lasers presently require low input
power and produce wider linewidths.
What is needed is a gas monitoring system that (1) allows simultaneous
determination of the concentration of components in a multiple component
gas; (2) is very efficient and uses as much of the total signal available
as possible for examination of the spectrum; (3) allows interrogation of a
broad spectrum in a single measurement of a small single sample; (4)
allows distinctions to be made relatively easily between gas components
that may be present; (5) allows considerable freedom in the choice of
sample container and sample concentration; (6) allows the system to
monitor and identify different compounds with minimal system
modifications; and (7) allows use of low power light sources and of
relatively low cost detectors.
SUMMARY OF THE INVENTION
These needs are met by the invention, which provides in one embodiment, a
light beam of a selected wavelength .lambda..sub.0, chosen to promote
Raman scattering in at least one of a predetermined group of gases that
may be present in the patient's airway; and a sample cell to receive and
temporarily hold gas from a patient's airway, the cell being relatively
transparent along the beam axis to light of wavelength .lambda..sub.0 and
being positioned to receive the light beam so that gas in the sample cell
is illuminated by the light beam. The invention also provides spectrograph
means, having a beam entrance and wavelength dispersion means, for
receiving a scattered light beam at the beam entrance, for directing the
scattered-and-received light beam to the wavelength dispersion means for
dispersing this light beam into a plurality of light rays, each ray having
a different interval of light wavelengths. Light collection means is also
provided, for collecting portions of the light beam that undergo Raman
scattering by one or more of the gases in the sample cell and for
directing these portions of the light beam to the spectrograph means beam
entrance. The invention further provides an array of photodetectors, each
of which receives one of the plurality of wavelength-dispersed light rays
from the spectrograph means and issues a photodetector output signal
representing the amount of Raman scattered light received by that
photodetector. Finally, the invention provides computation means for
receiving the photodetector output signal from each of M photodetectors
(M>number of gases to be tested for) and for computing the relative or
absolute amount of Raman scattered light received by each such
photodetector.
This invention measures the complete Raman spectrum of a gas mixture to
determine its composition. The invention gathers information about all the
gases present simultaneously, and from a single, small sample volume.
Further, the invention will combine techniques of light collection,
spectral separation, and signal processing to maximize the total amount of
light collected by the spectrometer, and to utilize the information
contained in the entire Raman spectrum to predict more accurately the
composition of the mixture. It is also the intention of this invention to
provide a means whereby new or additional gases can be detected in the
airway gas mixture, without altering the basic spectrometer, sample cell
or structure of the algorithm used.
The invention is an efficient means for collecting the optical signals
generated by Raman scattering of an optical beam in gases.
The Basic Invention Consists of the Following Essential Components
1. A monochromatic source of illumination, e.g., a laser;
2. A means of sampling and containing the gaseous media;
3. A system of collection optics;
4. A wavelength-selective filter to distinguish the Raman signals;
5. A photodetector; and
6. Computation means to determine gas composition from photodetector
signals.
Earlier inventors have avoided using a grating spectrometer as a
wavelength-selective filter because of high cost and low optical
efficiency. The invention described here combines several new ideas and
technologies that reduce these limitations. Traditionally, grating
spectrometers are used to observe the molecular structure of compounds.
These systems are often optimized for high resolution, and have low
optical throughput--a narrow entrance slit and "slow" optics. These
spectrometers are often designed to observe a wide range of compounds,
whose structure or spectra are often unknown.
In contrast, this invention employs a grating spectrometer to measure the
concentration, not molecular structure, of a limited number of known
compounds, such as the airway gases. This invention takes advantage of
these differences to optimize the combination of spectrometer optics,
unconventional entrance slits, and diode array elements for the
determination of airway gas concentrations. The optimized optical system,
in further combination with an intensified source of Raman-scattered
light, and an optimized data reduction technique, are part of the present
invention. Key elements of the new invention are use of 1) a means to
maximize the intensity of Raman-scattered light from a small aperture, 2)
a large etendue for the spectrometer, and 3) an optimized data reduction
means to determine concentrations.
BRIEF DESCRIPTION OF THE DRAWINGS
FIG. 1 is a schematic view of one embodiment of the invention.
FIGS. 2 and 3 are graphical views of the Raman scattering spectra for two
gases, N.sub.2 O and CO.sub.2, that are likely to be present in a
patient's airway.
FIGS. 4 and 5 are schematic views illustrating details of light
illumination apparatus for the gas samples held in the sample cells.
FIG. 6 is a schematic view illustrating use of a variable wavelength filter
to implement the invention.
FIGS. 7, 8, 9 and 10 are schematic views illustrating use of light-emitting
apertures with tubular sample cells according to other embodiments of the
invention.
DESCRIPTION OF BEST MODE OF THE INVENTION
A preferred embodiment 11, shown in FIG. 1, includes an optical waveguide
13, through which a gas sample 14 flows. A laser beam 15 passes along the
longitudinal axis of the waveguide 13 and produces Raman scattering along
the entire waveguide length. The laser beam 15 is preferably directed
through the waveguide 13 to minimize diffuse scattering of the laser beam
from the interior surfaces of the waveguide, but this is not essential.
The laser that produces the laser beam 15 may be a gas laser, such as
He-Ne or argon-ion, but is preferably a solid-state laser diode emitting
visible or near-infrared light in the wavelength range .lambda.=500-850
nm. One embodiment of the waveguide 13 is described in a co-pending patent
application by Carlsen et al, "System for Collecting Weakly Scattered
Electromagnetic Radiation", U.S. Ser. No. 07/840,108 filed Feb. 24, 1993,
and assigned to the assignee of this invention. A preferred embodiment of
the waveguide 13 is a cylindrical tube whose internal surfaces are
substantially fully reflective at the Raman shifted wavelengths of
interest, for example a smooth silver-plated surface approximately 100 mm
long with inside diameter between 1 and 2 mm. In this manner, the
effective length of interaction (dz in Equation (2)) is increased
substantially, thereby increasing the total Raman signal. In another
preferred embodiment, the laser beam propagates in an optically resonant
cavity that is surrounded by the waveguide 13.
The scattered light | | |
