header photo


Yale Astronomy

Iodine Cell Construction

Iodine cells are used for the wavelength calibration of a high-resolution RV spectra used for Doppler measurements in exoplanet searches. The absorption spectrum of molecular iodine provides a rich forest of lines from about 501-610 nm. The cell is wrapped with heat ribbon to keep it warm at the Observatory (so that the iodine is sublimated into a gaseous state) and a temperature probe and controller maintain the cell temperature above the temperature used to fill the cell with iodine vapor. The cell must be absolutely stable (sealed vacuum and non-reactive with the trace I2 gas) for timescales of decades to be useful for the detection of exoplanets with orbital periods of several years. Here we describe construction of ten iodine cells for our lab and spectrographs.

Figure 1 - Cell with manifold

Figure 1. The iodine cell before adding iodine: the middle section has a chamber for the iodine grain and the end of the manifold can be connected to vacuum for extracting water vapor and evacuating the cell.

We also require an extremely high resolution, high signal-to-noise spectrum of the iodine gas. In the past, the Fourier Transform Spectrograph (FTS) could be used at optical wavelengths at Kitt Peak, however this facility has been closed since about 2007. We identified two other FTS spectrographs that could be configured for optical wavelengths: one at NIST and one at the EMSL division of Pacific Northwest National Labs. Our FTS spectra have resolutions of about 1,000,000 and SNR of ~1000. We obtain FTS scans at different temperatures and interpolate between these to best match the spectra of our Iodine cells at the Keck, Lick and CTIO Observatories.


  1. quartz (fused silica) stock glass tubing for the body of the cylinder
  2. optically flat circular windows: 40mm Dia. 1/10 Wave Fused Silica Window, Uncoated (Edmund Optics Stock No. NT47-835)
  3. RTD temperature sensors (Minco part number: S665PDZ40AC)
  4. Kapton thermal ribbon (Omega) KH-206/10 POLYIMIDE FILM FLEXIBLE HEATER and KH-206-/10-P
  5. Heat resistant tape - Cole Parmer EW-08277-55
  6. pure grade iodine
  7. Temperature controller - Cole Parmer EW-89000-10
  8. melamine foam pipe insulation (McMaster 1.5” inner diameter and 2” inner diameter for the 40 mm and 50 mm diameter cells) - 93495K19
  9. TA3FX connectors for RTD sensors - Newark 89K7623
  10. SPT-1 cords and connectors for heaters - Digi-Key AE9903-ND
  11. temperature-regulated water bath

Construction of cells:

The cells were constructed by the glassblower in the chemistry department at Yale. Optically flat (l/10) windows were fused to the tube with sealing tape (G-1015, see http://www.vitta.com/gtapes.html) in an oven onto two sizes (40 or 50 cm diameter) of stock glass tubing. A tube was fused to the side of the cylinder and connected to a manifold for the iodine injection. Each cell has a unique identifier applied with heat tape by the glassblower.

Figure 2 - Drawing of I2 cell

Figure 2. Schematic drawing of iodine cell provided to the glassblower.

Iodine column density

The cells must contain a sufficient column density of iodine vapor for the wavelength calibration. If the column density is too low, then the absorption features will not be deep enough after the instrumental PSF from a typical high resolution spectrograph broadens the lines. However, if the column density is too large, then there will be an unnecessary loss of light. The column density is set by the vapor pressure and temperature of the water bath, used to sublimate the solid iodine grains.

AR coating for the optical windows:

We initially ordered optically flat windows with AR coating on both sides to minimize light losses. A test cell was constructed and put into the oven (to fuse the window onto the cylinder). When the cell came out of the oven, the AR coatings were cloudy and had to be discarded. Therefore the cells were constructed without AR coatings and the external windows will be solgel coated in our lab.

Figure 3 - Cloudy AR coating

Figure 3. The optical window with AR coating on both sides was cloudy and therefore unusable after being heated to about 900C in the glassblower oven. This high temperature was required for sealing the windows to the glass cylinders.


The cost includes materials, glassblower shop costs, heat ribbon, temperature probes, insulation, pure grade iodine and the cost of obtaining an FTS scan (contract to EMSL). Table 1 summarizes the costs.


After the cells were constructed by the glassblower, they were rinsed (through the tube on the side of the cylinder) with alcohol and then with de-ionized water until clean. The cells were put in an oven (900 degrees) overnight to dry. The manifolds were then attached by the glassblower and the cells were pumped to a ~10-4 vacuum (see Table 2, Vacuum w/o I2” for individual cells) to further extract any remaining water or vapor. A grain of iodine was dropped into the middle section of the manifold and the vacuum pump again applied (Vacuum w/ I2). The stopcock to the vacuum pump was closed and the cell and manifold were dunked into the water bath. Cells were filled at one of two different temperatures: 40C and 37C (water bath temperatures are recorded in Table 2) to fill the cells with more or less (respectively) iodine vapor. The cells all had a visible light pink vapor and cells filled at 40C had a darker pink-purple color (more iodine).

We inspected the cells (comments for individual cells in Table 2) and assigned a grade to each cell. Most of the cells had varying degrees of cloudy smudges on the cylinder walls – the glassblower thought this was “something that happened in the oven.” Two of the cells were clearly inferior and considered inadequate for use at an Observatory and therefore were not sent to EMSL for an FTS scan. One cell was left as a blank for calibrating the FTS. We taped the temperature probe near the middle of the cell cylinders and wrapped the cells with two Kapton heat ribbon. The Kapton’s were taped into place and wrapped with heat durable pipe insulation. The wires (to the Kapton’s and the temperature probe) are fed out through the pipe insulation. Each cell required two Kapton heat ribbons, and the wires had to be soldered for a parallel circuit connection and connectors attached.


TABLE 1: Project Budget

Item unit cost (US$) total (US$)
optical windows (20) 50 1000
Kapton heat ribbon 40 630
RTC temp probes 23.35 146.65
Temperature Controller 939.95 939.95
Heat resistant tape 41.5 41.5
Cords and connectors 138.97
Insulation (melanine pipe insulation) 33.6 67.2
Iodine 200
Glassblower labor and shop costs 3700
FTS scans (EMSL contract) shipping 8000


TABLE 2: Notes on Individual Cells

Cell ID Diameter [mm] Kapton resistance [Ohm] Vacuum w/o I2 Vacuum with I2 Water bath temp [C] Comments Cell Grade
A1 40 105 --- --- --- Good seals, clean windows. One window appears to be at a slant B
A2 40 105 1e-4 1.5e-4 --- Clean windows, good seals, flat window A
A3 40 105 4.8e-4 5.2e-4 37.2 Clean windows, flat, but one window slipped slightly. A-
A4 50 N/A --- 1.3e-4 --- Edges of window are ground down. One window has fairly significant slip. Possible stress lines run along the length of the cylinder. Spare for demos - will not be sent for FTS scan. B-
A5 50 N/A 1.6e-4 1.6e-4 --- Edges of window are ground down (reused optical flat windows). Both windows slipped. Ash or small spots inside the cell C
B1 40 105 --- 3.2e-4 --- Good seals. A
B2 40 220 3.4e-4 3.2e-4 39.7 One window slipped slightly. Dent in the cylinder where the evacuation tube was welded. B
B3 40 105 1.6e-4 1.6e-4 39.5 Good seals, slight dent in cylinder where tube attached A
B4 50 105 1.4e-4 3.2e-4 39.7 Good seal, ground edges A-
C 50 105 --- --- --- Clear cell - no I2. One window slipped slightly
C 50 105 --- --- --- Clear cell - no I2. One window slipped slightly
C 50 105 --- --- --- Clear cell - no I2. One window slipped slightly
1307a 40 --- --- 9.3e-05 41 Inner diameter 85 mm, slight wedge windows. AR dielectric coating from ECI. Dunk time: 2 min for the manifold and cell; this left residuals on the window that did not clean off completely and resulted in a very slightly degraded window quality. A-
1307b 40 --- --- 8.9e-05 40.5 85 mm inner diameter, wedge window. AR dielectric coating from ECI. Changed out the water bath, but still had some visible surface contaminants so only dipped the manifold with I2 into the water (2 min), leaving the cell in open air. Excellent window quality. A
1307c 40 --- --- 8.8e-05 40.5 85 mm ID, wedge window, AR dielectric coating from ECI. 2 min dunk of I2 manifold - cell left in air. Exellent window quality in central centimeter; partial ring compromises the window coating quality closer to the edge. A-
1307d 40 --- --- 8.8e-05 40.0 85 mm ID, AR coated from ECI, wedge window. Excellent window quality. 2 min dunk of the I2 manifold only. A
1307e 40 --- --- 8.7e-05 50 40 mm ID, AR coated from ECI, wedge window. Excellent window quality. 2 min dunk of the I2 manifold only. A
1307f 40 --- --- 8.4e-05 --- 85 mm ID, no coating, wedge window. Window distorted by vacuum (convex surface) during construction in glass lab; not useable as an I2 cell but saved as a blank for calibration of FTS (if useable). N/A
Figure 4 - Completed I2 cell  

Figure 4. The completed iodine cell with temperature probe taped to the cell wall, Kapton heat ribbon and melamie pipe insulation. All cells have a length of 11cm (for adequate column density of iodine). Cells were constructed with one of two diameters (40 mm cell diameter with 6.5 cm outer diameter including insulation; 50 mm cell diameter with 10 cm outer diameter including the insulation).