Patent Application
20080144007
This invention relates to Thermal Lens Spectroscopy (TLS) to provide ultra-sensitive absorption measurements, especially for the detection (i.e., identification) of analyte species and measurement of their concentration in micro-volumes of a carrier liquid.
Miniaturization has been a strong tendency in analytical systems development during at least the last decade. An important element of any analytical system based upon separation technologies is the ability to detect one or more analyte species present in low concentration in a sub-microliter volume of carrier liquid. The ability to detect a particular target analyte present at low concentration in a small detection volume becomes a critical issue especially in connection with micro-column high performance liquid chromatography (“HPLC”), and also in capillary electrophoresis (“CE”) where the detection needs to be done on an analyte sample present in a carrier liquid contained within a very small absorption cell, or within a CE capillary tube that frequently has diameter of 50 μm or less.
Traditionally, laser induced fluorescence (LIF) has been a sensitive analyte detection technique for micro-volumes due to its zero-background noise nature, and in some cases LIF permits detection of a single analyte molecule in the carrier liquid analytical volume. However, only a very limited number of analyte species have sufficient fluorescence efficiency, so fluorescent labeling or derivatization is necessary for most species. Additionally, the matrix effects associated with derivatization strongly impacts quantification capability, i.e., the ability to determine the target analyte concentration. An alternative method, namely ultra violet-Visible (“UV-VIS”) spectrophotometric absorption spectroscopy cannot provide sufficient sensitivity because of the optical path length inherent limitation of in-capillary detection. TLS, in accordance with the present invention, with its tightly focused probe beam, has intrinsic zero-background optical absorption measurement capability, and also extremely high sensitivity. TLS can therefore compete with LIF in sensitivity without its limitations as to detectable analytes.
a and 3b show two alternative relationships of the probe and excitation beams in conjunction with Thermal Lens imaging, also in accordance with the present invention.
a and 8b show a comparison of chromatograms obtained with a current state-of-the-art UV-VIS detector (Shimatzu SPD10AVvp), and a thermal lens detector in accordance with the current invention as shown schematically in
a and 9b show the comparative results obtained for 15 amino acids when using a prior art P/ACE-22-0 spectrometer and a TLS instrument in accordance with the present invention.
Although the basic principle of TLS were reported by M. Tokeshi, M. Uchida, A. Hibara, T. Sawada and T. Kitamori in “Determination of Submicromole Amounts of Nonfluorescent Molecules Using a Thermal Lens Microscope: Subsingle-Molecule Determination” Anal. Chem. 73, 2112-2116 (2001), this prior art reference provides neither an apparatus nor method possessing the advantages of the present invention which, depending on the particular embodiment selected, include:
An optical detector in accordance with the present invention is particularly suitable and advantageous for use in detecting concentration peaks on the output of a separation analytical tool such as HPLC (high performance liquid chromatography) or CE (capillary electrophoresis). Distinguishing features of my thermal lens detector are its very high sensitivity, ability to utilize an extremely small (sub-nL) carrier liquid volume, immunity to optical source noise and self-calibration of the time axis. My detector is based upon local change (i.e., change in the area of maximum excitation beam energy density) of the refractive index of the carrier liquid containing the target analyte sample thereby causing the carrier liquid to act as a microscopic thermal lens, which change is induced by the analyte's absorption of a focused beam (the “excitation beam”) at a pre-determined wavelength of interest (i.e., a distinctive absorption wavelength of the target analyte or analytes known or believed present in the carrier liquid). In operation the excitation beam is focused either into a quartz sample cell connected to a separation column when used in conjunction with HPLC, or directly into the capillary of a CE system, which capillary serves as an optical cell.
In HPLC the diameter of the focused excitation beam at the window of the cell containing the carrier liquid will preferably be not bigger than about 20 μm. The cell length will preferably be not bigger than about 100 μm, which results in a sample volume of less than about 50 pL (50×10−12 liter). When an absorbing species (in the case of either HPLC or CE separation) is present in the area of the cell or capillary illuminated by the excitation beam, the microscopic thermal lens, whose strength is proportional to the absorbed excitation beam power will change the divergence of a second laser beam, (the “probe beam”). This divergence change will result in a change in the optical power incident on a photo-detector having an aperture interposed between it and the source of the probe beam. This will, in turn, cause the photo-detector output signal to change, which change will be hereinafter referred to as the “thermal lens signal” or “TL signal”. The TL signal from the photo-detector is detected synchronously using, for example, a lock-in amplifier to thereby provide a signal which is proportional to the absorbed power of the excitation beam (which in a preferred embodiment will be at a UV wavelength of from about 200 to 350 nm). Where there is more than one target analyte, the excitation beam wavelength will be selected to coincide with an absorption wavelength of all the target analytes. From the time dependence of the TL signal the target analyte species can be identified using time coordinates, and the peak area or peak amplitude can provide the target analyte concentration. In
A graph illustrating the aforementioned time dependence is shown in
In TLS in accordance with the present invention, an excitation beam at a wavelength different from that of the probe beam (preferably having a wavelength shorter than the wavelength of the probe beam) is focused on a spot within, for example, an HPLC optical cell or a CE capillary channel (which capillary channel can be deemed an optical cell for purposes of this invention). The energy absorbed by the carrier liquid containing the target analyte (sometimes referred to as the “buffer” liquid) causes heating within a small region of the carrier liquid containing the target analyte, as explained in further detail below. The heated region is defined in the transverse direction by the focused excitation beam diameter (normally a few micrometers) and along the excitation beam axis the heated region will extend from the focal plane by about one Rayleigh length or less in the case where the beam path through the liquid contained in the cell is less than one Rayleigh length. In a preferred embodiment the excitation beam focus diameter will have a Rayleigh Range approximately equal to the length of liquid through which the beam passes along the beam axis. It is this local heating which creates a thermal lens due to the refractive index temperature dependence of the buffer liquid.
Such a thermal lens can be detected by a probe beam having a different wavelength which probe beam is also directed into the same sample volume. The divergence of the probe beam as a result of passing through the thermal lens will slightly increase, as is shown in
In traditional UV-VIS absorption spectroscopy one measures the absorbed fraction of the light intensity ΔI/I0≈1−10−A=1−e−αL≈αL, where I0 is the light intensity without absorption, α is the absorption coefficient, and A=cεL provides the absorbance of a solution of a sample of a length L with molar extinction ε and concentration c. In contradistinction, in TLS in accordance with the practice of the present invention, one measures the relative change of the probe beam signal intensity ΔI/I0 due to the thermal lens created by the excitation beam. This change depends on the excitation beam power PE, thermal conductivity k of the medium, (i.e., carrier liquid containing the target analyte), the probe beam wavelength λP and the temperature derivative of the medium refractive index dn/dT as:
One can thus see that the TLS signal is proportional to the sample absorption αL. The proportionality factor E:
can be referred to as the “Thermal Lens Enhancement Factor”, and indeed the relative change of the probe beam intensity can be larger than the classical absorption signal. For example, if one takes typical parameters for water, one finds that E=1 for a probe beam wavelength of 633 nm and an excitation beam power of only 4.2 mW. For excitation beam power values exceeding 4.2 mW the thermal lens signal will be significantly stronger than that of a classical absorption signal.
The excitation beam is preferably focused into the cell such that the beam's Rayleigh length (normally referred to as Zr) is chosen to be close to the cell length along the excitation beam propagation direction and its waist position is in the center of the cell. The probe beam is preferably focused to a diameter comparable to that of the excitation beam and the probe beam Rayleigh length will preferably be comparable to that of the excitation beam. However, the probe beam waist position, unlike the waist position of the excitation beam, will preferably not coincide with the excitation beam waist position but rather will be positioned several Rayleigh lengths (ranges) either behind or in front of the excitation beam waist. The Rayleigh length is defined as: Zr=f1w02n/λ where w0 is the beam waist radius, n is the refractive index of the medium containing the analyte and λ is the wavelength of the probe beam.
The overlap of the two beams is selected to give the maximum signal to noise ratio and hence the optimal thermal lens response. According to my invention, the beam axis of the probe beam is set at an angle to the axis of the excitation beam, such that the two beams overlap substantially or fully only within that part of the cell filled with the liquid containing the target analyte, but the two beams have at most limited overlap in the cell window. Such an arrangement of the excitation and probe beams is shown in
In some cases the reflection of the reference beam or of some part of it back to the probe beam source may occur, and such reflection can result in increased amplitude noise of the probe beam. In order to reduce the probability for such noise to occur, in an alternative embodiment of the present invention the probe beam and the excitation beam can both be advantageously directed along the capillary, i.e., the axis of both beams are situated in the plane that contains the capillary axis as is shown in
In another embodiment as shown in
In one preferred embodiment a transimpedance amplifier (TIA) is advantageously connected to the common point of the two photodiodes so that no current enters the TIA, thus producing no output signal when no thermal lens is present. The presence of a thermal lens will cause changes in the probe beam intensity transmitted through the aperture, while the intensity of the reflected portion of the probe beam will remain unchanged because it is solely focused onto the reference photodiode. In this balanced arrangement all fluctuations of the probe beam laser power are compensated for and thus do not provide any signal. Any noise on the signal detector is therefore due only to the combined shot noise of the probe beam photocurrent and reference beam photocurrent. The noise floor of the photo-detector is larger than the signal photocurrent noise by a factor of only √2. The relative noise value is therefore equal to √{square root over (4eBi0)}, where e is the electron charge, B is the detection bandwidth, and i0 is the average value of the signal diode photocurrent. If one assumes 4% of the total power in the signal beam and a 1 Hz detection bandwidth, then for a probe beam having 10 mW of optical power the value of the relative noise will be 8×10−8 which is more than three orders of magnitude better than the absorbance noise of existing, state of the art UV-Visible light commercial spectroscopic detection devices for use with CE or HPLC. This is a significant advantage of the present invention.
For a person skilled in optics it will be evident that a coating can be applied to the flat back surface of the off-axis lens in order to increase the fraction of the reflected probe beam that provides the reference beam and thereby further improve the signal to noise ratio of the photo-detector.
An alternative embodiment of the invention is shown in
In addition to the signal and reference photodiodes as shown in
If only absorption signals are detected, the identification of individual analyte species can be done by the time of appearance of their peaks (called “migration time” in CE). As the migration time is a function of the capillary temperature, the identification of an analyte by migration time alone requires temperature stabilization of the capillary to a small fraction of a degree, which becomes increasingly difficult the longer the capillary. The effect of the capillary temperature change is a substantial increase (or decrease) in the migration times for all analytes as a function of temperature. In order to compare two electropherograms taken at slightly different capillary temperatures one needs to re-adjust their time scales in order to make the peaks for the same species coincide. With the fluorescence signal, once the peak for a particular analyte has been identified, such time scale adjustment can readily be done. This permits one to relax significantly the requirements for the temperature stabilization of the capillary even to the extent of making temperature control unnecessary for some classes of analyte measurements.
In one embodiment of the current invention, the excitation light source is a pulsed high repetition rate laser. As an example, such source can be a diode-pumped solid-state laser having a saturable absorber inside its cavity, or a diode pumped optical fiber laser. The pulse repetition rate in such lasers can range from ˜30 to 40 kHz to ˜30 to 40 MHz, their average power can be in the range from a few mW to a few tens of mW resulting in peak power from a few hundred W to a few thousand W, and the individual pulse duration can be in the range of from a fraction of a picosecond to several nanoseconds. Such pulsed lasers are commercially available and have an affordable price and compact dimensions compared to CW lasers. The short pulse duration of such sources provides several advantages. A significant advantage of such pulsed sources, as is known in the art, is that nonlinear frequency conversion is more efficient, which makes it possible to obtain average powers from a few mW to several tens of mW at the fourth or even fifth harmonics, thereby bringing the excitation wavelength down to the deep UV range (≦˜300 nm). For amino-acid detection the wavelength of the excitation light source has a significant role, since most amino acids absorb only below 220 nm. In the case of such amino acids, I therefore prefer to use laser sources based upon nonlinear frequency conversion to provide excitation light of a UV wavelength below 220 nm.
In addition to this advantage, such a light source has an additional advantage in that the range of species that can be detected is broader because even species that have no linear absorption at the excitation wavelength will manifest two-photon absorption, and such species can therefore be detected by their two-photon absorption peaks, which would be invisible with a continuous-wave excitation light source. This is possible because a pulsed laser light source provides significantly higher (by several orders of magnitude) peak power. It is, of course, well known to the skilled artworker that two-photon absorption is proportional to the square of the peak power, e.g., a pulsed laser peak power 103 higher than a CW laser provides a two photon signal 106 stronger than for a CW laser of the same average power.
Even though there may be species that have no linear absorption even in the deep UV range, all analyte species show unique Raman spectra corresponding to its molecular vibrations. This feature provides the capability of detecting virtually any species. However, both two-photon and Raman signals can be more difficult to calibrate for quantitative measurements. In the present invention however, the TL signal in this instance provides the additional feature of providing intensity calibration using simultaneously Raman spectra and optical absorption as shown in
Besides the sensitivity enhancement, TLS in accordance with the present invention has the following important advantages:
All this makes TLS superior in accuracy even compared to other highly sensitive methods of absorption measurement such as Cavity Ring-Down Spectroscopy (CRDS). This can be seen from a comparison of the application of TLS to an HPLC separation column with the application of CRDS to the same system, as reported in K. Bechtel, R. Zare, A. Kachanov, S. Sanders and B. Paldus, “Moving beyond Traditional UV-Visible Absorption Detection: Cavity Ring-Down Spectroscopy for HPLC”, Anal. Chem. 77, 1177-1182 (2005). Such a comparison also illustrates an additional value of TLS system calibration in terms of measured absorption. CRDS is a direct absolute absorption method, in that it provides quantitative absorption data that can be used for this purpose. In the previous CRDS work, the same laser excitation light source at 488 nm (as is also suitable for the practice of the present invention) was used, which thus permits a direct comparison between classical UV-VIS absorption, CRDS and TLS. Five anthraquinone dyes which absorb in the blue spectral region were used in this experiment. The detection wavelength was 632.8 nm. The results and performance specifications in this experiment can be readily extended for other excitation wavelengths because of the universally applicable character of the thermal lensing effect.
The TLS potential can be understood from a consideration of
Substantially similar performance for TLS detection can be obtained for a fused silica capillary with a channel cross section of 50 μm, (an industry standard size in CE) based upon the fact that the actual probe volume size in another TLS measurement was 20 μm in diameter, i.e., significantly smaller than the 50 μm capillary channel size. When detecting a species in CE, the advantage of TLS will be even greater, as compared to HPLC. The very small capillary cross-section makes light collection from conventional small-area deuterium arc lamps less efficient than in HPLC, and therefore the baseline noise in CE devices with UV-VIS detection is higher than in HPLC. According to my measurements, the 3σ baseline noise in an industry standard CE instrument (Beckman P/ACE-2200) is 3×10−5 AU. Therefore the expected gain in performance of TLS versus UV-VIS in CE can reach three orders of magnitude.
This can be seen from back-to-back comparison of the measurements of 15 amino-acids made with Beckman P/ACE-2200 instrument and a TLS detection system in accordance with the present invention, the results of which are shown in
The list of amino-acids, their concentrations in the standard sample and in the diluted sample as well as the limits of detection (LOD) are presented in the Table 1. One can see that detection of amino-acids in nano-molar range are now possible with TLS.
| Number | Date | Country | |
|---|---|---|---|
| 60875035 | Dec 2006 | US |
