In order to overcome the limited capability of the high resolution etalon to separate between aerosol and molecular scattering and to increase the system stability and reliability, an iodine absorption filter was constructed. For the first HSRL measurements a 43 cm long cell was made. The cell was made from glass tubing with an attached side arm. Optical quality end windows with anti-reflection coatings were epoxed to the ends. The cell with iodine crystals in a side arm was evacuated and kept at 27 C. Transferring of the iodine from the absorption cell into the vacuum pump was prevented by evacuating the cell through a cold trap and cooling the side arm with liquid nitrogen. Although the iodine cell can be operated at room temperature, the operating temperature of the cell has to be controlled, because the vapor pressure of iodine is very temperature sensitive . In the HSRL, the cell temperature is maintained with 0.1 C accuracy by operating the cell in temperature controlled environment.
The iodine spectrum is measured by scanning the laser wavelength by changing the temperature of the seedlaser under computer control. A small amount of laser light is directed into a 100 m long fiber optic delay (Fiber 1 in Figure 4) and sent to the receiver to create a calibration light source. The temperature-wavelength dependence of the scan was determined by using the free spectral range of the high resolution etalon as a reference. This could be made, because the free spectral range of the etalon can be calculated when the length of the etalon spacers is known and the spacing of two (or more) etalon transmission peaks in temperature units can be measured. The calibration was made by simultaneously measuring the transmission spectrum of the high resolution etalon and iodine absorption filter. The simultaneous measurement of the high resolution etalon and iodine absorption filter transmissions was made by measuring the signal reflected from the high resolution etalon (Figure 12) and the signal transmitted through the absorption cell. The pressure in etalon was held constant while the laser wavelength was scanned. The spectrum was normalized by measuring the cell transmittance without the iodine cell.
A part of the measured iodine spectrum is presented in Figure 13. The measured spectrum was compared with a published spectrum  and an 0.01 pm wavelength agreement in relative line positions was observed. The linearity of the temperature scan was confirmed from the free spectral range information of high resolution etalon by performing the scan over more than one free spectral range. Single mode operation between two seedlaser mode hops can be maintained over 20 GHz range (at 1064 nm) and within this range two high resolution etalon free spectral ranges can be covered. During a mode hop the laser frequency jumps back about 10 GHz.
Figure 12: The HSRL receiver used for iodine spectrum calibrations. The same receiver setup was used for the first HSRL measurements with the iodine absorption filter. For data taking the transmission of the high resolution etalon was tuned out from the peak and the etalon was used as a reflector. When the beamsplitter and the mirror 4 are removed from the system, the system returns back to the old HSRL receiver.
Figure 13: Transmission of a 4 cm and 43 cm iodine cells as a function of wavelength shift. The identification line numbers are from Gerstenkorn and Luc.
For initial HSRL measurements the line 1109 (peak wavelength 532.26 nm), which is well isolated from the neighboring lines, was chosen. The full width half maximum width of the line is 1.8 pm and the peak transmission is 0.08%. The hyperfine structure of the peak 1109 defines the asymmetric shape of the absorption peak . In fact, the line 1109 is a combination of two rotational vibrational transitions with different hyperfine structures.
The iodine absorption cell provides a robust filter for the HSRL, because it is not dependent on the mechanical alignment of the filter or the angular dependence of the incoming light. Another advantage is the stability of the absorption characteristics. This provides a stable long term operation. The strength of observed absorption line is dependent on the line strength, and the length, temperature, and pressure of the cell. By controlling the operating environment and with a nearly leak proof system, the current iodine cell is operated for several months without any maintenance. During this time, a small change in absorption strength and line width were observed due to a small leak that was caused by the iodine penetrating through a hose. Also the iodine was found to condense into the walls of the cell, but even during a long period of time, the amount of condensation has been small and 10% extra absorption is observed. The condensation can be prevented by operating the tip of the side arm couple degrees below the cell temperature. The problems with reactive iodine penetrating through the hoses can be prevented by using a sealed all-glass cell. In a short term operation, the stability of the absorption characteristics has proven to be so good that a system calibration scans from different days can be used for the calculations of the system calibration coefficients. This requires, that the alignment of the receiver optics is stable.
An absorption filter offers a high rejection against aerosol scattering and therefore it makes the separation between aerosol and molecular scattering easier. Also, a wide dynamic range in rejection against aerosol scattering is achieved by simply changing the vapor pressure or the length of the cell. Comparison between high resolution etalon and iodine absorption filter performance is presented in Figure 14. A 2:1 separation between molecular and aerosol scattering by the etalon (Figure 14.b) is measured compared to a 1000:1 separation in the iodine cell when operated at 27 C (Figure 14.a). The molecular transmission in Figure 14.a and Figure 14.b is calculated by using the Doppler-broadened molecular spectrum at -65 C. This temperature is close to the lowest temperature measured at the tropopause and this gives the smallest transmission through the iodine absorption cell. The molecular transmission of the high resolution etalon and the iodine absorption filter are similar (Figure 14.c). Due to wide absorption line width, the molecular transmission of the iodine filter is more dependent on the air temperature than the etalon. The temperature dependence of the cell transmission is modeled by using the table values of iodine vapor pressure  (Figure 14.d). Calculations show, that by changing the cell temperature from 27 C to 0 C, the online transmission can be tuned from 0.08% to 60%.
Figure 14: (a) Transmission of 43 cm cell (solid line) together with the molecular transmission (dashed line) at -65 C air temperature as a function of wavelength shift. Dot-dashed line shows the calculated molecular spectrum at -65 C. (b) Etalon transmission (solid line) and calculated molecular transmission (dashed line) as a function of wavelength shift. Dot-dashed line shows the calculated molecular spectrum at -65 C. (c) Comparison of molecular transmission of high resolution etalon and iodine cell as a function of air temperature. (d) Iodine cell aerosol and molecular transmission as a function of cell temperature.