Acoustic Wave H.sub.2 O Phase-Change Sensor Capable of Self-Cleaning and Distinguishing Air, Water, Dew, Frost and Ice

1. Field of the Invention

This invention relates to devices used for frost sensing in a "demand-defrost" refrigeration system. The device can be applied broadly to heatabsorbing surfaces and to measure surface condensation from the atmosphere. Typical heatabsorbing surfaces to which the sensor may be applied include an evaporator tube, fin or shell in a refrigerator or heat pump. Another typical surface on which surface condensation and accretion detection are important is an aircraft wing.

2. Description of the Prior Art

Previous frost sensors include capacitive, optical, and acoustical structures. Capacitive structures experience an increase in capacitance due to frost deposition between plates monitored. Optical modules experience attenuation or scattering of light due to frost deposition. Acoustical cavities or waveguides experience an attenuation of sound due to frost deposition.

All three frost detection methods have drawbacks. There is a lack of discrimination between frost and ice in capacitive or optical schemes. The lack of thermal contact to the evaporator may limit the accuracy of optical devices. The size of acoustic waveguides limits their direct placement within available spaces in an evaporator.

It is an object of this invention to create a frost sensor of small size which may be fabricated in batches such as integrated circuit chips. The small size allows a fit into closely spaced openings such as between fins of an evaporator for direct sensing of cold surfaces.

It is another object of this invention to create a frost sensor capable of discriminating between air, water, dew, frost, and ice. The frost sensor should also discriminate between frost thicknesses. The device should exhibit large signal changes, for relative ease of H.sub.2 O phase-change determination. Discrimination between H.sub.2 O phasechanges will allow shortening of a refrigeration defrost cycle.

It is another object of this invention to create a frost sensor with a hydrophobic surface. The hydrophobic surface will allow self-cleaning during defrost and protection from contaminants.

The present invention relates to a discovery that surface acoustic wave (SAW) devices exhibit high sensitivity to frost/ice deposition and to liquid/solid phase changes of H.sub.2 O. The present invention consists of a SAW phase-change sensor with short acoustic path length on a strong piezoelectric substrate. The sensor can be batch fabricated, and its small size allows placement in good thermal contact with a cold surface such as a refrigeration evaporator fin. The SAW phase-change sensor experiences large changes in insertion loss and a.c. impedance upon deposition of H.sub.2 O in various phase changes. The sensor has a hydrophobic surface treatment and placed in a vertical position will self-clean during a refrigeration defrost cycle.

Further details of the present invention are explained with reference to the attached drawings in which:

FIG. 1A shows a plan view of a SAW sensor made in accordance with the invention;

FIG. 1B shows a cross section of the SAW sensor of FIG. 1A taken along line A--A' thereof;

FIG. 2 illustrates the self cleaning orientation of the SAW phase-change sensor with hydrophobic surface treatment to repel liquid water;

FIG. 3A shows the mounting of the SAW phasechange sensor of FIG. 1A in a simulation apparatus; and

FIG. 3B shows a block diagram of test equipment configuration for testing the test setup of FIG. 3A.

FIG. 1A shows the SAW phase-change sensor 1 of the present invention. FIG. 1B shows a cross sectional view of the sensor of FIG. 1A taken along line AA' of FIG. 1A. The thicknesses in FIG. 1B are not to scale. The SAW phase-change sensor 1 consists of a piezoelectric substrate 5, patterned metallization 8, a thermally conductive package 11, thermally conductive adhesive 13, wirebonds 15, and hydrophobic surface treatment 17.

As an example, a batch of sensors 1 was fabricated with X-propagation on a 128 degree rotated Y-cut lithium niobate substrate (128Y-X LiNbO.sub.3). The lithium niobate thickness was 0.5mm. The metallization 8 was RF-diode sputtered Cr/Au (50.ANG./1000.ANG.), and was patterned using conventional photolithography. The metallization patterns formed interdigitated transducers (IDT's) which were conventional .lambda./4 types with period .lambda.=52.5.mu.m and an aperture width w=2800.mu.m. This yields a center frequency of approximately 75 MHz. One IDT serves as an input transducer indicated at 9 whereas the other IDT serves as an output transducer indicated at 10. As an example, the input IDT may have 10 finger pairs and the output IDT 30 finger pairs. Center-to-center spacing of the two IDTs was 30.lambda.(1575.mu.m). This spacing is quite small compared with many SAW designs, but is advantageous for lossymedium sensing applications because the short path length results in manageable total insertion loss. The ratio of substrate thickness to acoustical wavelength is typically 9 or higher so the Rayleigh propagation mode tends to dominate. After metallization and oscillation checks for every metallization pattern, a wafer which consists of approximately 80 metallization patterns was diced.

Once diced, the individual sensors 1 were individually mounted in gold-plated nickel thermally conductive packages, or flatpacks 11. The piezoelectric substrates were attached to flatpacks 11 using thermally conductive epoxy adhesive 13. The metallization patterns were then wirebonded at point 15 to make electrical connections to the flatpack. Finally, a hydrophobic surface treatment 17 was formed on the surface of each sensor by treating the metallization patterns with a polysiloxane derivatizer.

FIG. 2 demonstrates the self cleaning orientation of the SAW phase-change sensor 1 with hydrophobic surface treatment to repel water, droplets of which are indicated at 19. The hydrophobic behavior is particularly advantageous when the sensor surface is oriented vertically, as demonstrated in FIG. 2 by the gravity direction lines 18. Being oriented vertically in simulated frost/defrost cycles, the water 19 which formed from melting frost or ice was drained from the sensor surface due to gravity 18 and the preferential wetting of surrounding surfaces. This effect is advantageous because the sensor tends to be self-cleaning when liquid is present, such as during a defrost cycle.

Even with the sensor mounted horizontally, the liquid will wet the flatpack surfaces, but will be repelled from the surface of sensor 1 by the hydrophobic surface treatment 17. During experimentation, 0.5 ml of distilled water added to the flatpack wetted the flatpack interior, but was repelled from the substrate 5 and formed an air bubble arching over the surface of substrate 5 which persisted about 10 s before collapsing.

FIG. 3A illustrates a simulation apparatus used in testing the sensor 1 and is seen to include a Peltier thermoelectric cooler 20 for simulating an evaporator cold surface and a thermocouple 22 for measuring surface temperature. A SAW phase-change sensor 1 is mounted in thermal contact with one face of the Peltier thermoelectric cooler 20. Thermocouple 22 is placed in thermal contact with the flatpack 11.

FIG. 3B shows a block diagram of the test equipment configuration for testing the setup of FIG. 3A. A network analyzer 30 is connected to the SAW phase-change sensor 1 to measure transmission (insertion loss) and input reflection (impedance). A digital thermometer 32 is connected between the thermocouple and network analyzer 30. The network analyzer 30 also measures analog voltages from the digital thermometer 32 which provides voltage signals proportional to thermocouple 22 temperature. Surface temperature was varied by supplying dc power from a regulated power supply 34 to the Peltier thermoelectric unit 20. A vapor source 36 was provided to generate humid air. The vapor source 36 operated by bubbling compressed air through water and directing this saturated air stream toward the chilled sensor when dew or frost generation was required.

Table 1 summarizes initial measurements taken to confirm that the SAW phase-change sensor could distinguish between air, dew, water, and frost. Measurements were made using the configuration of FIG. 3B. Measurements were made with the SAW phasechange sensor surface horizontal. Transmission was measured from 65 MHz to 85 MHz with the minimum insertion loss and its corresponding frequency listed in Table 1. Table 1 tabulates the measurement results.

The first measurement for air shows an insertion loss of 9.2 dB at 74.7 MHz. Filling the flatpack with distilled water and chilling to -2.degree. C. increased the insertion loss to 28.4 dB at 74.6 MHz. There is minimal shift in frequency of the peak, probably because only a very thin layer of water couples to the 75 MHz SAW. The supercooled water persists in liquid form for many minutes before freezing.

When the dry sensor is chilled in air, dew formation occurs to an extent dependent on the ambient dewpoint. There is no appreciable effect on insertion loss until the sensor surface temperature approaches the dewpoint. Insertion loss of the sensor due to dew deposited by chilling to -2.degree. C. is 21.7 dB at 74.4 MHz. As is the case for bulk liquid water, supercooled dew persists for many minutes at -2.degree. C. When the dew freezes, the acoustic response of the sensor is dramatic. With almost identical surface coverage of frost the insertion loss changes to 36.0 dB and the peak frequency shifts to 75.4 MHz, a substantial change for a SAW device.

Table 2 shows a compilation of multiple measurements made with SAW phase-change sensors fabricated as discussed herebefore to confirm that the SAW phase-change sensor could distinguish between air, dew, frost, and water. The insertion loss measurements in Table 2 were made using SAW phase-change sensors over the stated temperature range for each phase-change condition.

If the insertion loss measurement is made at fixed frequency within the SAW passband, the sensor's response to frost and ice can be more easily distinguished.

Table 3 shows the ability of the SAW phasechange sensor to distinguish various frost thicknesses, and to distinguish between frost and ice.

With the vapor source off and a tight glass cover on the flatpack to prevent condensation, the dry sensor was chilled to -4.degree. C. The insertion loss at 74.0 MHz was 8.2 dB.

With the vapor source off and the flatpack open to room air, the dry sensor was chilled to -4.degree. C. A thin noncontinuous layer of dew grew on the sensor surface and then froze. The layer thickness of frost was approximately 0.05 mm. The insertion loss at 74.0 MHz was 40.4 dB.

The vapor source was turned on to grow a thicker frost layer. Growth was continued at -4.degree. C. to reach an approximate layer thickness of 0.5 mm before the vapor source was turned off. The insertion loss increased to 48.9 dB at 74 MHz.

More frost was deposited at -4.degree. C. using the vapor source until approximate thickness was 2 mm. The insertion loss increased to 52.5 dB at 74 MHz.

An ice layer about 3 mm thick on the sensor was formed by filling the flatpack with distilled water and chilling quickly to -4.degree. C. The insertion loss was 59.2 dB at 74 MHz. For comparison, a bulk water layer about 3 mm thick was obtained by filling the horizontal flatpack to form a large meniscus. By slow chilling, the water was supercooled down to -4.degree. C. and the sensor insertion loss was 30.2 dB at 74 MHz.

It can be seen that, at constant temperature, the sensor's insertion loss rapidly increases with frost deposition and then approaches a limiting value as frost thickness increases. The insertion loss due to ice is still about 6 dB larger than that from frost. The insertion loss due to water is at least 10 dB less than that from frost.

An alternative to insertion loss is the use of electrical reflection measurements from a single IDT of the SAW phase-change sensor. Either input impedance Z* (.omega.) or input admittance Y* (.omega.)=1/Z* (.omega.) can be obtained, with admittance being preferred for SAW device input properties. Y* is a complex quantity Y*=Y'+Y"; one can obtain either the real or imaginary part with the identifications Y' (.omega.)=G (.omega.) and Y"(.omega.)=.omega.C(.omega.), where G (.omega.) is "radiation conductance" and C (.omega.) is "electrostatic capacitance" of the input IDT at frequencies within the SAW passband. The quantity .omega. is 2.pi. times frequency.

Input measurements of G (.omega.) and C (.omega.) for a SAW phase-change sensor were made with various water deposits. At suitable fixed frequency, the radiation conductance G (.omega.) is very sensitive to the water phase-change, as shown by Table 4.

Table 4 shows that separation between frost and bulk ice using impedance measurement is much wider than is the case for the insertion loss measurements. In addition to these large changes in input conductance at fixed frequency, the frequency peak value changes greatly. For example, maximum conductance shifts from 72.7 MHz (air loading) to 72.3 MHz (frost); to 71.4 MHz (bulk water); and to 70.3 MHz (bulk ice).

The invention has been described above with particularity so as to teach one skilled in the art how to make and use the invention. Many modifications will fall within the scope of the invention. For instance:

Any non-corroding, highly conductive, adherent, low-contact-resistance, thermal-expansion compatible, bondable metallization system useable with thin-film deposition, patterning and etching to required IDT dimensions can be substituted for Cr/Au.

Any chemically inert piezoelectric substrate capable of crystallographic orientation, surface finishing and withstanding photolithographic processing can be substituted for the 128Y-X LiNbO.sub.3. However, the strong piezoelectric coupling coefficient (K.sup.2 =0.052) and low acoustic loss (.alpha..about.0.01 dB/cm-MHz) of this material result in SAW sensors with a low unloaded insertion loss (7 to 9 dB at 75 MHz), high sensitivity to deposited mass and hence large changes in insertion loss with frost deposition. The added insertion loss due to frost loading should roughly scale linearly with K.sup.2, other factors being equal.

Any hydrophobic surface treatment compatible with the sensor materials may be used, although that used in the disclosed sensor contributes no measurable insertion loss and is thus preferred.

The acoustic path length may be varied in order to adjust the total insertion loss with frost/ice/water loading. Reduced path length would reduce loaded insertion loss roughly linearly with length, other factors being equal but also would reduce sensitivity to phase-change and introduce substantial acoustical reflections between IDT's.

The IDT period .lambda. may be altered in order to adjust operating frequency to lower or higher values, which may affect loaded insertion loss sensitivity.

Several other thermally conductive package configurations may be chosen. However, a flatpack allows excellent thermal contact to be achieved between the sensor and the cold surface, along with excellent corrosion resistance, electrical and mechanical stability and low parasitic capacitance. Electrical connection to the thermally conductive package may be obtained by means other than wirebonds.