Open Access
Add to BibSonomyAbstract
Current atomic clocks are burdened by size, weight, power and portability limitations to satisfy a broad range of potential applications. One critical need in the fabrication of a miniaturized atomic clock is small, low-power metallic sources. Exploiting the relatively high vapor pressure of ytterbium (Yb) and its dissolution in anhydrous ammonia, we report two independent techniques for depositing Yb inside a well micromachined into a microhotplate. Subsequent in situ evaporation of Yb from the microhotplate well serves as a low-power metallic source suitable for atomic clocks. The deposition and evaporation of Yb were confirmed using a variety of physicochemical techniques including quartz crystal microbalance, scanning electron microscopy, energy dispersive X-ray spectroscopy, and laser fluorescence. We also describe the fabrication of the microhotplate device, an integral component of our Yb-based miniature atomic clock. The Yb deposition/evaporation on a microhotplate well is thus useful as a low power Yb source during the fabrication of a miniaturized atomic clock, and this technique could be used for other applications requiring a vapor of a metal that has a moderate vapor pressure.
©2012 Optical Society of America
1. Introduction
Atomic clocks have numerous already-realized and potential applications ranging from the sophisticated to the critical but mundane. Such applications include commercial and military communication and navigation systems, surveillance, weapons guidance, data encryption, managing electrical grids, remote sensing as well as fundamental studies in atomic physics [1]. Currently available commercial atomic clocks are based upon the hyperfine splitting of an alkali atom, and the international standard for the second is defined relative to the hyperfine splitting of the Cs atom. The Cs fountain clock (NIST-F1) keeps time to an accuracy of one second in one hundred million years [2]. However, high accuracy atomic clocks [3–6] are too large (several cubic meters), expensive and power hungry (upwards of 500 W to 1 kW) in order to satisfy the aforementioned applications. Even higher performing atomic clocks are being developed that are based on optical frequency transitions in atoms and ions where higher frequency laser light can provide finer temporal resolution [7–9]. While a great deal of research has been devoted to the improvement of atomic clock stability and accuracy, many practical applications can tolerate moderate performance from an atomic clock, but need high reliability and portability, requiring much reduced size, weight, and power.
In the last decade, there has been a significant effort toward the extreme miniaturization of atomic clocks (largely funded by the Defense Advanced Research Projects Agency (DARPA)). Much of this effort was devoted toward the miniaturization of vapor-cell-based atomic clocks, where a cell having a size on the order of 1 mm3 contains a vapor of either cesium (Cs) or rubidium (Rb) and a buffer gas. These miniature atomic clocks have been dubbed chip-scale atomic clocks (CSACs). They are characterized by small size (1-5 cm3), low power consumption (~100 mW) and a fractional frequency instability better than 10−11 at one hour [10,11]. Efforts in this area were led by, amongst others, Symmetricom [10,12], Honeywell [10], National Institute of Standards and Technology (NIST) [13], Sarnoff [14], and Teledyne [15]. More recent development has focused on drastically improving the long-term frequency stability (~10−14 at one month) of miniaturized atomic clocks while approaching the small size and power consumption characteristic of the CSACs. In this context, we are developing an atomic clock utilizing trapped ytterbium-171 (171Yb) ions by probing its 12.6 GHz hyperfine transition under the Integrated Micro Primary Atomic Clock Technology (IMPACT) project sponsored by DARPA [16]. The Yb trapped ion clock offers several advantages: availability and stability of isotopes, low cost relative to other rare earth metals, low nuclear spin (I = 1/2) in the 171Yb isotope, the relative accessibility of the 369 nm 2S1/2 to 2P1/2 transition for laser excitation, and a favorable characteristic of all trapped ion clocks, excellent isolation of the trapped ions from the surrounding environment such that their internal energy levels are not perturbed for optimum clock performance.
One critical need for a miniature atomic clock is a small, low power metallic source to provide an ensemble of atoms for the atomic clock. A chief hurdle is obtaining a purified source of metal that, when activated, will not disturb the atmosphere surrounding the atomic ensemble contained in a vacuum package or a vapor cell. We describe for the first time the dissolution of Yb in liquid anhydrous ammonia for its in situ deposition into a well micromachined into a microhotplate (hereafter referred to as microhotplate or microhotplate well, due to the contours of the device). Alternatively, the higher volatility of Yb relative to other rare earth elements [17,18] can be used to deposit a thin film of Yb into a microhotplate well. Once Yb is thus deposited, it can subsequently be released into the vapor cell of an atomic clock using the microhotplate’s inherent heating capability. Yb deposition by either of these techniques and its subsequent evaporation from a microhotplate well were confirmed using physicochemical techniques including mass changes by quartz crystal microbalance (QCM), scanning electron microscopy (SEM), energy dispersive X-ray (EDS) spectroscopy and the characteristic fluorescence emission from Yb. We also describe the fabrication of a microhotplate device, an important component of our overall miniature 171Yb trapped ion clock.
The miniaturized Yb source described in this correspondence may find uses with other types of atomic clocks and other applications requiring Yb. The National Institute of Standards and Technology (NIST) Yb optical clock is also based on Yb atoms (cooled to near absolute zero and trapped in an optical lattice) [19]. Ytterbium is one of the more commonly used ions in trapped ion quantum computation experiments. Such experiments frequently utilize integrated chip-based traps for sophisticated manipulation of the ions [20,21]. However, macro scale sources are used with the chip-based ion traps, and a miniature Yb source as discussed in this paper could be readily integrated with the chip-based traps making for a more compact system. With Yb, it is also possible to obtain a Bose-Einstein condensate for basic physics studies [22].
2. Experimental section
Microhotplate Fabrication and Characterization: The microhotplate structures were built from silicon-on-insulator (SOI) wafers. The top layer of the wafer, or device layer, is a highly doped 10 μm thick layer of p-type single crystal silicon with a resistivity of 0.005 to 0.02 ohm-cm. Deep-Reactive-Ion Etching (DRIE) was used to define the electrical conduction pathway in the device layer and to create the three-dimensional (3D) Yb-holding well structure on the backside of the 400 μm thick SOI handle layer. A layer of positive photoresist was lithographically-patterned in both procedures and was used to define etched (unmasked by the photoresist) and un-etched (masked) regions. For the device-layer etch, the mask and process were designed to completely remove the device layer down to the 1 μm thick insulating buried oxide (BOX) layer of the SOI everywhere except the current conduction pathway and the two bond-pad zones, positioned at each end of the conductor, for macroscale electrical connection. The exposed BOX appears green in Fig. 1(a) , while the silicon current conduction pathway is gray. The silicon bond pad zones were covered with 200 nm aluminum (containing 1 wt. % silicon) by an evaporative liftoff process. If required for subsequent packaging, the aluminum can also be covered with 10 nm titanium and 200 nm gold, by evaporative liftoff. Figure 1(a) provides a solid model of a dual-beam cantilever design of the microhotplate. The backside etch (Fig. 1(c)) created the roughly circular well needed to hold the subsequently-deposited Yb metal.
Fig. 1 Images of a completed microhotplate. (a) Solid model of the front side of a microhotplate. The yellowish ‘L’-shaped regions are where the aluminum bond pads are located. (b) SEM with scale bar (200 μm) from a working distance (WD) of 19.6 mm, and an EHT of 20.00 kV, showing the front surface. (c) Backside SEM taken at 45° to illustrate the thickness of Yb-holding well (volume, 0.707 mm3).
The backside etch also created a vacuum gap in order to reduce the thermal conduction out of the heating zone (Fig. 2(a) ). Double-clamped beam design microhotplates were also produced (Fig. 2(a)), but the out-of-plane expansion of the cantilever design (Fig. 1) offered better mechanical stability at high temperature. Since the emphasis was on mechanical stability, we did not perform thermal imaging on double-clamp beam design microhotplates.
Fig. 2 (a) Photograph of a microhotplate well which is 1.5 mm in diameter and 0.4 mm deep (height); b) Infrared microscope (IR) image of a powered microhotplate showing its thermal scan.
By placing conductive heating elements on a thin, free-standing structure such as these (Figs. 1 and 2), high temperatures can be rapidly achieved with low electrical input power. Microhotplates, as these devices have come to be known, are capable of reaching temperatures of 400-600°C in 1-20 msec with only 50-100 mW of power [23,24]. The aluminum bond pads were deposited on the device layer and sintered in forming gas (3% H2 in N2, 30 min., 425°C) in order to allow Ohmic electrical contacts to be made to the structure/device. Joule heating is capable of producing a sustained, stable temperature in excess of 600°C in the central heating zone. The temperature distribution of an electrically-heated microhotplate was measured using an infrared microscope (EDO Barnes Infrascope, Shelton, CT). The results are shown in Fig. 2(b). Using a two-temperature radiance technique [25], the Infrascope first generates an emissivity map on the surface of the device under test (DUT). Then, the radiance of the powered DUT was measured and a corresponding temperature map was calculated [25] (Fig. 2(b)).
Ytterbium: Yb (99.9% pure) was obtained either as tiny metallic bars (Fisher Scientific, Fairlawn, NJ) for liquid ammonia dissolution or as < 5 mm evaporation pieces (Kurt J. Lesker Company, Pittsburg, PA, part # EVMYBX5MM-D) for electron-beam (e-beam) experiments. Although, Yb is less expensive compared to other rare earth metals at ~$155/kg, isotopically-pure 171Yb is expensive at $14,000/g, available as foils, and there are only few suppliers for this material. As shown in this study (see below), an advantage of the anhydrous ammonia technique is that it is highly compatible with such expensive foils. For this reason, proof-of-concept experiments and optimization of Yb deposition and evaporation were carried out using the non-isotopically pure Yb samples mentioned above. Yb was kept in a nitrogen glove box inside a clean room and purged with nitrogen in order to minimize oxidation.
Liquid Ammonia Dissolution/Evaporation: We exploited the in situ dissolution of Yb in anhydrous liquid ammonia and its subsequent evaporation from a microhotplate well as one technique to supply the low-power metallic source. Here, the Yb was dissolved using liquid anhydrous ammonia and deposited into the microhotplate well. This brought the Yb into intimate contact with the microhotplate surface, a situation ideal for subsequent thermal evaporation of the metal, thereby reducing power consumption of the completed device, as well as confining the expensive metal rigidly inside the microhotplate well for future use. The chamber containing the microhotplate and Yb was ~15 mL and was filled with anhydrous ammonia to atmospheric pressure. A small spatula wrapped with a fabric covered wire was cooled in liquid N2. The microhotplate/Yb was cooled by applying this spatula to the outer wall of the reaction vessel until liquid ammonia began to condense on the microhotplate, at which time the cold spatula was removed. Not all of the ammonia in the chamber was condensed. When excess ammonia was condensed by prolonged cooling, the Yb coated both the outside of the microhotplate as well as the chamber glass. It is likely that a metered amount of ammonia in the chamber, followed by cooling to condense all the ammonia present, would provide a more reproducible process and increase the yield. With practice, this qualitative cooling process was reproducible and fairly high yielding. After Yb deposition, the microhotplate could be heated or vacuum pulled on the containment vessel to evaporate the liquid ammonia, leaving behind a pure metal slug in the microhotplate well. Anhydrous liquid ammonia dissolution of Yb is shown schematically in Fig. 3 .
Fig. 3 Schematic of Yb dissolution in anhydrous liquid ammonia. The reaction vessel was evacuated and back-filled with anhydrous ammonia. Brief (5 sec) cooling of the vessel under the device condensed a small amount of ammonia and reflowed the Yb. The condensed ammonia was allowed to evaporate in the closed system. After one minute, the excess ammonia was removed under vacuum and the metal deposition was confirmed by examining the microhotplate well using a VistaVision light microscope.
Electron-Beam (e-Beam) Yb Thin Film: We reported that higher melting point metals such as platinum (Pt) (1772°C) could be evaporated from a microbridge heater device [26]. We therefore exploited the relatively lower melting point (824°C) of Yb to enable thin film deposition into and subsequent re-evaporation from a microhotplate to provide a different, independent route for creating a low-power Yb source. We used a Temescal FC/BJD200 (Livermore, CA) system for e-beam evaporation of Yb into the microhotplate with a dedicated tungsten crucible to prevent potential cross-contamination. The base pressure for evaporation of a 2 μm thin film/layer of Yb onto the microhotplate was 6 x 10−7 Torr. The Temescal power to evaporate Yb was 1.8% of full scale, demonstrating low-power consumption compared to other metals such as Pt, which is often done at 70% of full-scale power. Individual microhotplates were clamped to the evaporator platen under a shadow mask which allowed metal deposition only in the microhotplate wells. Shadow masking also helped to thermally isolate the active zone by minimizing thermally-conductive metallic layers from depositing on the microhotplate suspension beams.
Microhotplate Test Assembly: We assembled the microhotplate into a custom-built vacuum chamber operated at ~10−8 Torr. We fabricated a vacuum-compatible circuit board in aluminum nitride (MicroFab Inc., Manchester, NH) to interconnect a Yb-coated microhotplate (from liquid ammonia or e-beam deposition methods) to a heater control circuit. The microhotplate was epoxied to the circuit board, and electrical connections from the hotplate to the board were made with aluminum wirebonds. A low-outgassing card edge connector (Sullins, Inc., San Marcos, CA; model # WMC05DTEH) joined the circuit board traces to solder cup connectors. Vacuum-compatible, insulated 22 gauge copper wire was attached to the solder cups using 91% Sn, 9% Zn solder (H & N Electronics, California City, CA). The coated device and board were inserted into an instrumented vacuum chamber which also contained a quartz crystal microbalance (QCM) (Inficon, East Syracuse, NY; model # CDS-A0F47, SQM-160 controller) aligned directly above the microhotplate (line-of-sight), a residual gas analyzer (RGA) (Stanford Research Systems, Sunnyvale, CA, model # RGA200) and an optical port for measuring laser fluorescence. Laser fluorescence was excited with a 399 nm laser resonant with the 1S0 to 1P1 transition in neutral Yb. The 399 nm light is generated by doubling ~30 mW of light from a distributed feedback (DFB) laser (Eagleyard Photonics, Berlin, Germany, model # EYP-DFB-0795-00080-1500-TOC03) at 798 nm. The laser was frequency doubled to 399 nm using a periodically-poled KTP (potassium titanyl phosphate) waveguide crystal (ADvR Inc., Bozeman, MT) producing ~100 μW of light. The 2 mm diameter laser beam was directed across the microhotplate well and brought to a focus ~1 cm above it using a 30 cm focal length lens, while an imaging system and photon multiplier tube (PMT) counter (Hamamatsu H7155, Bridgewater, NJ) with a laser line filter (Semrock Inc., Rochester, NY; model # FF01-395/11-25) to reject ambient light were placed perpendicular to the laser beam path. The microhotplate was controlled by a programmable DC power supply and overall experimental control and data collection was done with LabView software. The test assembly is shown in Fig. 4 .
Fig. 4 Yb deposition test assembly. The microhotplate Yb source vacuum mount is shown on the top left and after it is inserted into the vacuum chamber on the top right. The final location of the vacuum chamber in the test assembly is indicated by the arrow. The various measurement instruments (QCM, PMT, RGA) are also indicated.
Yb Evaporation: The microhotplate loaded with Yb was inserted into a vacuum system (Fig. 4), and then the Yb was evaporated from the well using the integrated heater (Fig. 2(b)). The basic method was simple. The voltage to the microhotplate under observation was stepped up at a known rate to a preset maximum voltage. Once the peak voltage was reached, it was held at that value for the duration of the experiment. Once the soak time had elapsed, the power supply was set to zero volts and the experiment was concluded. Data was collected throughout the entire ramp and soak. The following data were collected: photon count (0.01 s gate), quartz crystal microbalance (QCM) frequency, applied voltage, actual device voltage and the voltage across a low-value (10.2 ohm) series current-viewing resistor. All other parameters were calculated, such as device resistance and device temperature. With regard to the latter, an Agilent 6890 oven (Agilent Technologies, Santa Clara, CA) was used to measure the microhotplate resistance, R, against the temperature, T, from 60° to 275°C. The data was fit to the function,
with a coefficient of correlation, r2 of 0.999. The resistance versus temperature calibration is expected to be valid until the intrinsic temperature of the silicon conduction layer is approached, which is well above the evaporation temperature used in these studies. This equation can be used to estimate the average temperature of the microhotplate well, and is expected to provide a good estimate of the average temperature of the thin-film Yb layers, precisely because they are very thin, thermally conductive and in excellent contact with the surface. The actual temperature of the liquid-ammonia-deposited Yb, however, is overestimated by Eq. (1) due to its relative thickness and non-uniform contact with the microhotplate well. This formula is also sensitive to possible variations in the baseline resistance value.Yb Characterization: We used several different physicochemical techniques to evaluate Yb deposition and evaporation. The QCM utilizes the piezoelectric sensitivity of a quartz crystal in order to measure mass changes. When mass is added to the crystal its resonant frequency is reduced, thus, Δf = fq – fc, where, Δf is the frequency change, and fq and fc are the frequencies of the uncoated and coated (mass added) quartz crystal frequencies, respectively. This change in frequency is highly sensitive and can detect the addition of less than one atomic layer of mass [27]. The mass of material Δm added to the QCM is calculated by
where, A is the exposed area of the QCM (0.49 cm2), fo its natural resonant frequency (5.974 MHz) and ρ and μ the density (2.65 g.cm−3) and shear modulus (2.95 x 1011 g.cm−1.s−2), respectively, of the QCM substrate. Scanning electron microscopy (SEM) images and data were collected using Zeiss Supra 55VP field-emission SEM (Carl Zeiss AG, Germany) equipped with Bruker Quad Silicon Drift (SD) detector and Bruker Esprit acquisition software (Bruker AXS Corporation, Madison, WI). Energy Dispersive X-ray Spectroscopy (EDS/EDX/EDAX) was used for elemental analyses such as hydrogen, nitrogen, oxygen, and of course, Yb. Since each element has a unique atomic structure, the X-rays produced are also unique for a particular element. A Zeiss 55VP field-emission SEM with a Bruker Quad SD detector was used for EDS analysis with Bruker Esprit software for data acquisition and mapping. Details of QCM, RGA and fluorescence measurements are described above.3. Results
Figure 5 shows the SEM and EDS images of Yb deposited onto microhotplate using the anhydrous liquid ammonia technique. With EDS, we are detecting the presence of Yb based upon the energy (in keV) of the emitted X-rays shown on the abscissa. This is distinct from the detection of fluorescence from optically excited Yb vapor (discussed elsewhere in the paper).
Fig. 5 SEM and EDS results of the microhotplate coated with Yb using the anhydrous liquid ammonia technique. The EDS tracings show the presence of Yb. The black-and-white inset is the SEM image of the microhotplate showing Yb deposition. The blue inset shows the same Yb deposition by EDS. The magnification was 43X, the filament potential was 15.0 kV, and working distance (WD) was 14.9 mm from the SEM pole piece (focal length). The scale bar is 500 μm for both inset images.
The EDS results (Fig. 5) showed that pure Yb metal and metal oxides remained in the device well whereas ammonia, nitrogen and hydrogen were undetectable, demonstrating the complete evaporation of liquid ammonia post-deposition and the absence of nitride phases. The liquid ammonia technique adhered Yb metal to the microhotplate well and allowed a greater quantity of Yb to be deposited as compared to the e-beam technique.
When the heater was activated by a current, the microhotplate was heated to ≥ 600°C using 1 W of power on a ~1 mm3 plate well resulting in a cloud of Yb atoms being released from the Yb layer reaching the QCM. Typical results are shown in Fig. 6 . The voltage to the microhotplate was ramped in a controlled way to a constant set point, producing a square-wave temperature response. We measured Yb fluorescence, QCM frequency change, chamber pressure, and microhotplate temperature as a function of elapsed time (Figs. 6 and 8 ). The QCM, positioned above the microhotplate, initially shows a frequency increase, which is interpreted through Eq. (2) as a mass decrease. The apparent cause of this anomaly is the heat radiated from the microhotplate or the temperature of the deposited Yb. As the temperature of the microhotplate stabilizes, so does the rate of change of mass deposited on the QCM. When the microhotplate was turned off, the sign of the anomalous mass change was reversed, supporting our theory that the temperature was influencing the QCM response. We estimated the Yb evaporation rate from the slope of the QCM response, once this response had stabilized. This method provides a rate of 23 pg/sec of metal being deposited on the QCM. Assuming an average atomic mass of 173 g/mol, this represents an atomic rate of 8.0 x 1010/second arriving at the QCM. The average temperature of the microhotplate well bottom was determined from its measured voltage and current and Eq. (1). The average mass of Yb deposited in microhotplate wells by the liquid ammonia technique was 209 ± 56 μg (n = 3). Had the microhotplate well been filled to capacity, we would estimate a Yb mass of ~4.9 mg based upon the well volume of 0.707 mm3 and a density of Yb of 6.9 g/cm3. Because of the relatively large absolute thickness of the Yb deposited by the anhydrous liquid ammonia technique, and the changing thickness of the Yb as a function of elapsed time, the actual evaporation temperature of Yb was not determined. However, the temperature of the microhotplate well bottom is provided to illustrate the time during which it was heated in situ in order to produce Yb evaporation (Fig. 6).
Fig. 6 In situ evaporation of Yb from a microhotplate well. The Yb was deposited into the microhotplate well using anhydrous liquid ammonia. The heat pulse caused an apparent drop in mass at first, then the QCM frequency change properly indicated Yb evaporation. The measurement parameters are color coded and numbered as follows: microhotplate temperature, red tracing, # 1; PMT counts per 10 ms, black tracing # 2; QCM response, blue tracing, # 3.
Fig. 8 Responses from QCM, fluorescence and temperature, indicating the evaporation of Yb from the microhotplate well. For this experiment, thin film evaporation was used to deposit the Yb metal inside the microhotplate well. The measurement parameters are color coded and numbered as follows: microhotplate temperature, red tracing, # 1; PMT counts per 10 ms, black tracing, # 2; QCM response, blue tracing, # 3.
The characteristic Yb fluorescence appeared above the microhotplate after the heat was applied (Fig. 6). The laser is tuned to the optical resonance of ytterbium-174, the most abundant isotope. Systematic uncertainties in the Yb fluorescence measurement made it difficult to ascertain quantitatively the Yb flux with a high degree of confidence, but it does illustrate well how the relative flux changes with time. The fluorescence signal in Fig. 6 shows an initial increase as the voltage is ramped up over 140 s to 13.5 V, slowly rises over time, and then is ramped off over 13 s starting at 4298 s. We confirmed that the measured signal is not due to black body radiation by tuning the laser frequency off the 174Yb resonance line and observing a count rate equal to the background count rate of ~1200 per 10 ms. There is a distinct jump up in the fluorescence at ~3300 s and back down at 4100 s. This is due to the laser “mode hopping” and thereby slightly changing its frequency and output power. As the heater is ramped on and off, the fluorescence is at 10% of the nominal “fully on” value found at 13.5 V. This seemingly sudden switching on and off of the evaporation is due to the Yb vapor pressure varying exponentially with the temperature, so small changes in temperature have a large effect on the evaporation rate. In addition, the Yb vapor does not persist in the vacuum chamber because it has a very high probability of adsorption on the room-temperature chamber surfaces—no Yb is observed in the RGA during evaporation. However, we do see the total pressure rise during the heating of the microhotplate, which we attribute to the components surrounding the hotplate slowly heating up and outgassing.
Figure 7 contrasts the physical appearance of the Yb, as-deposited by the liquid ammonia technique, with that after a microhotplate evaporation step like that of Fig. 6. The SEM images before and after evaporation indicate a redistribution of Yb in the well. The regions surrounding the microhotplate, such as the top edge of the isolation structure, appears brighter post evaporation due to the deposition of Yb. The EDS data post evaporation indicates that the chemical composition of the material in the microhotplate well is similar to that of the as-deposited Yb, as inferred from Fig. 5, which is typical for all the as-deposited Yb samples measured.
Fig. 7 SEM and EDS data for a microhotplate with liquid-ammonia-deposited Yb. The central SEM shows the Yb as-deposited. The remaining data were taken after evaporation of the Yb from the microhotplate. Details are similar to those in the legend to Fig. 5.
Similar experiments were performed on the Yb-deposited in the microhotplate well by thin-film evaporation (Fig. 8). As with the liquid ammonia-deposited Yb, the pressure in the chamber and fluorescence were observed to rise with the microhotplate temperature. With thin film Yb deposition however, the microhotplate temperature was expected to be much more representative of the Yb evaporation temperature since the Yb thin film was considerably more uniform and only 2 μm thick at the beginning of the experiment. Using the dimensions of the microhotplate, we can estimate the mass of 2 μm Yb at ~10 μg.
The mass of Yb deposited on the QCM, as calculated from the measured QCM frequency change (Eq. (2), is also shown in Fig. 8. While the microhotplate temperature was steady, the QCM mass change rate was also steady at a value of approximately 9.6 pg/sec of Yb being delivered to the QCM (Fig. 8).Since the angle subtended by the QCM is a small fraction of the total solid angle for evaporation from the microhotplate, the actual rate of Yb evaporation would be considerably higher than the calculated 9.6 pg/sec. Stated differently, the 9.6 pg/sec represents the lower limit of Yb evaporation from the microhotplate, when it was deposited as a thin film. Following this experiment, the microhotplate was examined by SEM and EDS in a manner similar to those represented in Figs. 5 and 7. The data, not shown due to manuscript space limitations, illustrates that the thin film Yb was completely removed from the microhotplate by evaporation.
The QCM response during the evaporation of Yb deposited using the anhydrous liquid ammonia technique, shown in Fig. 6, is qualitatively similar to the thin film Yb evaporation data of Fig. 8. This response included the anomalous sign changes during the microhotplate heating and cooling and a steady change while the microhotplate temperature was stable. The Yb deposited using the anhydrous liquid ammonia technique was relatively thick and, as mentioned above, delivered an average rate of 23 pg/sec (8.0 x 1010atoms/second) of metal to the QCM once the QCM response had stabilized. Similar to the thin film evaporation, this value too represents only the lower limit for the total evaporation rate given the solid angle arguments (see above). The resulting flux of Yb was determined to be satisfactory for a working atomic clock. It should be noted that the total capacity of the well in a typical microhotplate (1.5 mm diameter, 0.4 mm deep) is approximately 4.9 mg, though this amount can be increased or decreased depending on the well size and the clock power constraints.
4. Discussion
Using MEMS (MicroElectroMechanical System) techniques, we demonstrated the fabrication of a microhotplate device suitable for use with miniaturized atomic clocks. The MEMS approach offers several critical advantages to miniaturized atomic clocks in designing and developing an ultra-small physics package [1]. MEMS has enabled a high level of component miniaturization, lower cost, high volume robust and reproducible fabrication [11,28–31] as well as monolithic (or hybrid) integration [32] allowing for significant reductions in size, weight, power and portability (SWAPP) without compromising the performance metrics and perhaps actually improving upon them.
Experimental evidence from other Yb sources used in our lab indicates that a microhotplate filled using the anhydrous ammonia technique should provide years of operation of an atomic clock provided care is taken to use the minimum evaporation rate necessary to load the ion trap. Due to the long ion storage times associated with ion traps, on the order of 1000 hours [33], a continuously operating clock would need to load only monthly to maintain the required number of ions in the trap. Typical ion loading times range from tens of seconds to tens of minutes depending on the neutral atom flux and the ionization technique used. It is likely that as the clock is miniaturized, the total amount of Yb required would be decreased since the Yb source would be closer to the ion trap so that a larger fraction of the emitted flux would pass through the ion trap. Only neutral atoms passing through the ion trap can be ionized and subsequently trapped. In this case one can consider smaller capacity Yb sources which would lead to lower power consumption and lower overall size. Further refinements to miniaturization may employ free-standing micro-machined filaments known as microbridges [26,34]. A typical microbridge is 10 μm wide, 2 μm thick and is elevated above the substrate by an air gap of approximately 2 μm. Length scales of the microbridge are in the range of 100 μm to 1 mm. In addition, microbridges will outperform microhotplates in terms of operational power (35 mW) and a response time of 0.2 msec compared to 20 msec for the microhotplate. Finally, to ensure a high percentage of 171Yb ions in the trap, isotopically-enriched Yb will be used in these devices.
We employed two independent techniques for the deposition and subsequent evaporation of Yb metal from the device. Rare earth elements including Yb are susceptible to oxidation. We have found that the thin films of Yb can be in air for several days, and we are still able to observe Yb evaporation. Evaporation in this manner onto the microhotplate well could turn wasteful since the surface area of the microhotplate is very small compared to the interior of an evaporation chamber or a sputtering chamber. This inefficiency can be ameliorated to some extent by placing an array of several microhotplates on a coater platen or deposit on to a wafer of devices before dicing the wafer in order to reduce waste. As commented in the Experimental section (above), waste would be a significant concern with the isotopically-pure Yb. In this context, the liquid ammonia deposition technique (Fig. 3) is more advantageous.
The mechanism of the liquid ammonia technique is as follows: Anhydrous ammonia is a Lewis basic high polarity solvent [35]. Low concentrations of Yb in liquid ammonia are blue in color whereas more concentrated solutions assume a copper-bronze hue presumably due to the formation of an electride salt. The blue color is due to the solvated electrons which exist alongside the metal cation [35]. In this context, the term “solvated electrons” refers to the electrons in solution or electrons trapped within the solvent cavity. Solvation must therefore be conceptually distinguished from dissolution, which is a kinetic process characterized by rate. Sufficient energy must be released during Yb dissolution in liquid ammonia in order to overcome the sum of the heat of sublimation (ΔHs) and the ionization potential (IP). Yb has a relatively low IP (39.2 eV) and ΔHs (4.41 eV). Therefore, its charge density, Z2/r (solvation energy of the metal ion plus the electron solvation energy), need not be large in order to compensate for the former factors (IP + ΔHs) [35–37]. These properties explain the facile dissolution of Yb in liquid ammonia and its subsequent deposition into the microhotplate well.
Finally, we compare and contrast our Yb deposition techniques with those published previously for alkali atom filling of atomic clock vapor cells. There are indeed a number of different techniques that have been used for filling atomic clock vapor cells with metal atoms [1]. For example, several papers were published based on the reaction between alkali metal chloride and barium azide for the production of alkali metal atoms in order to fill vapor cells [10,29,30,38–41], thus:
BaN6 + MeCl → BaCl + 3N2 + Me, where Me represents the alkali metal
The metal atoms used included Cs and Rb. With BaCl being left behind in the vapor cell as a dark colored solid, there is the potential for Ba and N2 to recombine to form BaN6 which could deplete the N2 buffer gas. Other techniques for the deposition of alkali atoms included thin film and ultraviolet decomposition of CsN6 to produce pure Cs and N2 [31]. Gong et al [42] used Cs-enriched glass as a source of in situ production of Cs atoms through electrolytic decomposition at 500°C. Finally, Rb encased in wax micropackets was used to deposit the metal after laser ablation at 200°C [43]. The aforementioned techniques for introducing alkalis into miniature atomic vapor cells are addressing the fact that the alkalis oxidize quickly when exposed to air and the deposition and the sealing of the vapor cell must occur in an oxygen- and water-free environment. However, these techniques introduce substances into the vapor cells that may affect long-term clock performance. We have found that the miniature Yb sources can be exposed to air for several days to weeks without seeing an adverse amount of oxidation. Our microfabrication and deposition processes are generally applicable to other elements used in atomic clocks and quantum information processing. These include the alkali and alkali earth metals and Cd, but apparently not aluminum [35,36,44]. The rapid oxidation of the alkali metals would complicate the use of the microhotplates with the alkalis.
5. Conclusions
The technologies described for the first time in this paper enable a miniature, low-power Yb source with a supply of Yb adequate for long-term operation, especially with the use of the liquid ammonia technique. Miniaturization of the Yb source was achieved using microfabrication processes for manufacturing micro-containers with integrated heaters. The dissolution (reflow) of comparatively low melting point/volatile Yb in liquid ammonia enabled its in situ deposition in a microhotplate. Successful deposition of Yb using liquid ammonia or e-beam thin film techniques and subsequent evaporation of Yb were confirmed using physicochemical methods. These sources are intended for use in an Yb trapped ion atomic clock, and their inclusion in a miniature ion trap/vacuum package is currently being tested.
Sandia is a multiprogram laboratory operated by Sandia Corporation, a Lockheed Martin Company, for the United States Department of Energy under Contract DE-AC04-94AL85000. This work was supported by the Defense Advanced Research Projects Agency (DARPA) micro-PNT program IMPACT effort under agreement # 017081218, which DARPA has Approved for Public Release, Distribution Unlimited. We thank Bonnie McKenzie for SEM and EDS analyses. The views expressed are those of the authors and do not reflect the official policy or position of the Department of Defense or the U.S. Government.
Acknowledgments
References and links
1. S. Knappe, “MEMS atomic clocks,” in Comprehensive Microsystems, Vol. 3 (Elsevier, 2007), 571-612.
2. T. E. Parker, “Long-term comparison of caesium fountain primary frequency standards,” Metrologia 47(1), 1–10 (2010). [CrossRef]
3. N. D. Lemke, A. D. Ludlow, Z. W. Barber, T. M. Fortier, S. A. Diddams, Y. Jiang, S. R. Jefferts, T. P. Heavner, T. E. Parker, and C. W. Oates, “Spin-1/2 optical lattice clock,” Phys. Rev. Lett. 103(6), 063001 (2009). [CrossRef] [PubMed]
4. C. W. Chou, D. B. Hume, J. C. Koelemeij, D. J. Wineland, and T. Rosenband, “Frequency comparison of two high-accuracy Al+ optical clocks,” Phys. Rev. Lett. 104(7), 070802 (2010). [CrossRef] [PubMed]
5. J. J. McFerran, L. Yi, S. Mejri, S. Di Manno, W. Zhang, J. Guéna, Y. Le Coq, and S. Bize, “Neutral atom frequency reference in the deep ultraviolet with fractional uncertainty=5.7 x 10−15,”Phys. Rev. Lett. 108(18), 183004 (2012). [CrossRef] [PubMed]
6. N. Huntemann, M. Okhapkin, B. Lipphardt, S. Weyers, C. Tamm, and E. Peik, “High-accuracy optical clock based on the octupole transition in 171Yb+.,” Phys. Rev. Lett. 108(9), 090801 (2012). [CrossRef] [PubMed]
7. Ch. Lisdat, J. S. Winfred, T. Middelmann, F. Riehle, and U. Sterr, “Collisional losses, decoherence, and frequency shifts in optical lattice clocks with bosons,” Phys. Rev. Lett. 103(9), 090801 (2009). [CrossRef] [PubMed]
8. M. Petersen, R. Chicireanu, S. T. Dawkins, D. V. Magalhães, C. Mandache, Y. Le Coq, A. Clairon, and S. Bize, “Doppler-free spectroscopy of the 1S0-3P0 optical clock transition in laser-cooled fermionic isotopes of neutral mercury,” Phys. Rev. Lett. 101(18), 183004 (2008). [CrossRef] [PubMed]
9. X. Baillard, M. Fouche, R. Le Targat, P. G. Westergaard, A. Lecallier, F. Chapelet, M. Abgrall, G. D. Rovera, P. Laurent, P. Rosenbusch, S. Bize, G. Santarelli, A. Clairon, P. Lemonde, G. Grosche, B. Lipphardt, and H. Schnatz, “An optical lattice clock with spin-polarized 87Sr atoms,” Eur. Phys. J. D 48(1), 11–17 (2008). [CrossRef]
10. M. A. Perez, U. Nguyen, S. Knappe, E. A. Donley, J. Kitching, and A. M. Shkel, “Rubidium vapor cell with integrated Bragg reflectors for compact atomic MEMS,” Sens. Actuators A Phys. 154(2), 295–303 (2009). [CrossRef]
11. S. Knappe, P. D. D. Schwindt, V. Gerginov, V. Shah, L. Liew, J. Moreland, H. G. Robinson, L. Hollberg, and J. Kitching, “Microfabricated atomic clocks and magnetometers,” J. Opt. A, Pure Appl. Opt. 8(7), S318–S322 (2006). [CrossRef]
12. R. Lutwak, P. Vlitas, M. Varghese, M. Mescher, D. K. Serkland, and G. M. Peake, “The MAC – A miniature atomic clock,” in Proceedings of 36th Annual Precise Time and Time Interval (PTTI) Systems and Applications Meeting, L. Breakiron, ed., 752–757 (2005).
13. S. Knappe, V. Shah, P. D. D. Schwindt, L. Hollberg, J. Kitching, L. A. Liew, and J. Moreland, “A microfabricated atomic clock,” Appl. Phys. Lett. 85(9), 1460–1462 (2004). [CrossRef]
14. A. M. Braun, T. J. Davis, M. H. Kwakernaak, J. J. Michalchuk, A. Ulmer, W. K. Chan, J. H. Abeles, Z. A. Shellenbarger, Y. Y. Jau, W. Happer, T. McClelland, H. Fruehauf, R. Drap, W. Weidemann, and M. Variakojis, “Re-interrogated end-state chip-scale atomic clock,” in Proceedings of 39th Annual Precise Time and Time Interval (PTTI) Meeting, Long Beach, CA, 233–248 (2007).
15. J. F. DeNatale, R. L. Borwick, C. Tsai, P. A. Stupar, Y. Lin, R. A. Newgard, R. W. Berquist, and M. Zhu, “Compact, low-power chip-scale atomic clock,” in Position, Location and Navigation Symposium, 2008 IEEE/ION, 5–8 May, 67–70, (2008).
16. P. D. D. Schwindt, Y. Y. Jau, H. Partner, D. K. Serkland, R. Boye, L. Fang, A. Casias, R. P. Manginell, M. Moorman, J. Prestage, and N. Yu, “Micro ion frequency standard,” Proc. SPIE 8031, 803100 (2011).
17. A. H. Daane, D. H. Dennison, and F. H. Spedding, “The preparation of samarium and ytterbium metals,” J. Am. Chem. Soc. 75(9), 2272–2273 (1953). [CrossRef]
18. E. J. Onstott, “The preparation of samarium metal with calcium,” J. Am. Chem. Soc. 77(3), 812–813 (1955). [CrossRef]
19. Z. W. Barber, C. W. Hoyt, C. W. Oates, L. Hollberg, A. V. Taichenachev, and V. I. Yudin, “Direct excitation of the forbidden clock transition in neutral 174Yb atoms confined to an optical lattice,” Phys. Rev. Lett. 96(8), 083002 (2006). [CrossRef] [PubMed]
20. D. Stick, W. K. Hensinger, S. Olmschenk, M. J. Madsen, K. Schwab, and C. Monroe, “Ion trap in a semiconductor chip,” Nat. Phys. 2(1), 36–39 (2006). [CrossRef]
21. J. M. Amini, J. Britton, D. Leibfried, and D. J. Wineland, “Microfabricated chip traps for ions,” Atomic Chips (Wiley-VCH, 2012).
22. T. Hong, C. Cramer, E. Cook, W. Nagourney, and E. N. Fortson, “Observation of the 1S0-3P0 transition in atomic ytterbium for optical clocks and qubit arrays,” Opt. Lett. 30(19), 2644–2646 (2005). [CrossRef] [PubMed]
23. R. E. Cavicchi, J. S. Suehle, P. Chaparala, K. G. Kreider, M. Gaitan, and S. Semancik, “Micro-hotplate gas sensor,” Tech. Digest 1994, Sol. State Sensor Actuator Workshop, Hilton Head, SC, 53–56 (1994).
24. R. P. Manginell, D. R. Adkins, M. W. Moorman, R. Hadizadeh, D. Copic, D. A. Porter, J. M. Anderson, V. M. Hietala, J. R. Bryan, D. R. Wheeler, K. B. Pfeifer, and A. Rumpf, “Mass-sensitive microfabricated chemical preconcentrator,” J. Microelectromech. Syst. 17(6), 1396–1407 (2008). [CrossRef]
25. J. McDonald, “Microthermal imaging in the infrared,” http://www.Electronics-Cooling.com/Resources/EC_Articles/JAN97/jan97_04.htm; http://www.latigooptics.com/pubs&seminars/1997_microthermal_imaging.PDF.
26. R. P. Manginell, J. H. Smith, A. J. Ricco, D. J. Moreno, R. C. Hughes, R. J. Huber, and S. D. Senturia, “Selective, pulsed CVD of platinum on microbridge gas sensors,” Tech. Digest 1996, Solid State Sensors Actuators Workshop, Hilton Head, SC, 53–56 (1996).
27. S. S. Narine and A. J. Slavin, “Use of the quartz crystal microbalance to measure the mass of submonolayer deposits: Measuring the stoichiometry of surface oxides,” J. Vac. Sci. Technol. A 16(3), 1857–1862 (1998). [CrossRef]
28. S. Knappe, V. Velichansky, H. G. Robinson, J. Kitching, and L. Hollberg, “Compact atomic vapor cells fabricated by laser-induced heating of hollow-core glass fibers,” Rev. Sci. Instrum. 74(6), 3142 (2003). [CrossRef]
29. L. A. Liew, S. Knappe, J. Moreland, H. Robinson, L. Hollberg, and J. Kitching, “Microfabricated alkali atom vapor cells,” Appl. Phys. Lett. 84(14), 2694–2696 (2004). [CrossRef]
30. S. Knappe, V. Gerginov, P. D. D. Schwindt, V. Shah, H. G. Robinson, L. Hollberg, and J. Kitching, “Atomic vapor cells for chip-scale atomic clocks with improved long-term frequency stability,” Opt. Lett. 30(18), 2351–2353 (2005). [CrossRef] [PubMed]
31. L. A. Liew, J. Moreland, and V. Gerginov, “Wafer-level filling of microfabricated atomic vapor cells based on thin-film deposition and photolysis of cesium azide,” Appl. Phys. Lett. 90(11), 114106 (2007). [CrossRef]
32. R. P. Manginell, J. M. Bauer, M. W. Moorman, L. J. Sanchez, J. M. Anderson, J. J. Whiting, D. A. Porter, D. Copic, and K. E. Achyuthan, “A monolithically-integrated μGC chemical sensor system,” Sensors (Basel Switzerland) 11(7), 6517–6532 (2011). [CrossRef]
33. D. Prestage and G. L. Weaver, “Atomic clocks and oscillators for deep-space navigation and radio science,” Proc. IEEE 95(11), 2235–2247 (2007). [CrossRef]
34. C. Vauchier, D. Charlot, G. Delapierre, and A. Accorsi, “Thinfilm gas catalytic microsensor,” Sens. Actuators, B. 5(1–4), 33–36 (1991). [CrossRef]
35. J. J. Lagowski, “Solution phenomena in liquid ammonia,” Pure Appl. Chem. 25(2), 429–456 (1971). [CrossRef]
36. J. C. Warf and W. L. Korst, “Solutions of europium and ytterbium metals in liquid ammonia,” J. Phys. Chem. 60(11), 1590–1591 (1956). [CrossRef]
37. K. Howell and L. L. Pytlewski, “The decomposition products of solutions of europium and ytterbium metals in liquid ammonia,” J. Less Common Met. 19(4), 399–404 (1969). [CrossRef]
38. S. Knappe, V. Velichansky, H. G. Robinson, L. Liew, J. Moreland, J. Kitching, and L. Hollberg, “Atomic vapor cells for miniature frequency references,” in Proceedings of 2003 IEEE International Frequency Control Symposium and PDA Exhibition Jointly with the 17th European Frequency and Time Forum, Institute of Electrical and Electronics Engineers, New York, 31–32 (2003).
39. S. Knappe, P. D. D. Schwindt, V. Shah, L. Hollberg, J. Kitching, L. Liew, and J. Moreland, “A chip-scale atomic clock based on 87Rb with improved frequency stability,” Opt. Express 13(4), 1249–1253 (2005). [CrossRef] [PubMed]
40. E. J. Eklund, A. M. Shkel, S. Knappe, E. Donley, and J. Kitching, “Spherical rubidium vapor cells fabricated by micro glass blowing,” IEEE International Conference on Micro Electro Mechanical Systems (MEMS 2007), Kobe, Japan, 171–174 (2007).
41. E. J. Eklund and A. M. Shkel, “Glass blowing on a wafer lever,” J. Microelectromech. Syst. 16(2), 232–239 (2007). [CrossRef]
42. F. Gong, Y.-Y. Jau, K. Jensen, and W. Happer, “Electrolytic fabrication of atomic clock cells,” Rev. Sci. Instrum. 77(7), 076101 (2006). [CrossRef]
43. S. Radhakrishnan and A. Lal, “Alkali metal-wax micropackets for chip-scale atomic clocks,” Digest Tech. Papers of Transducers 05, 23–26 (2005).
44. E. W. Brooman, “Electrodeposition from liquid ammonia solutions,” Electrodep. Surf. Treat. 2(1), 1–46 (1973). [CrossRef]
