Comparison of both MEMS Thermal Conductivity CO₂ Sensors

And Non-dispersive Infrared CO₂ Sensors

Xiang Zheng Tu

 

A major hazardous gas present in the atmosphere which creates various adverse effects to human is carbon dioxide (CO₂). Measurement of CO₂ gas using CO₂ sensor will help to monitor its presence and indicate us above dangerous limits to prevent adversities. In a modern ventilation system CO₂ sensors used as indoor air quality indicators help to ensure a fresh outside air supply to building occupants while simultaneously optimizing energy consumption. For such systems it has been recommended: CO₂ sensors shall be certified by the manufacturer to be accurate within plus or minus 75 ppm at a 600 and 1000 ppm concentration. Carbon dioxide air-conditioning systems have been installed in cars. In these applications the CO₂ leakage detection range is 0-50.000 ppm with a resolution of 10% while the comfort range is 0-5.000 ppm with a resolution of 200 ppm.

A non-dispersive infrared (NDIR) CO₂ sensor is shown in the above first figure. The NDIR sensor comprises an infrared source, a sample cell, an optical filter and a detector system. The infrared source directs waves of light through the cell filled with air containing CO₂ toward the filter and then detector which measures the amount of the light that hits it. As the light passes through the cell, any gas molecules that are the same size as the wavelength of the light absorb the light only, while letting other wavelengths of the light pass through. Next, the remaining light hits the filter that absorbs every wavelength of light except the exact wavelength absorbed by CO₂. Finally, the detector reads the amount of light that was not absorbed by the CO₂ molecules or the optical filter.

The difference between the amount of light radiated by the source and the amount of the light received by the detector is measured. The difference is proportional to the number of CO₂ molecules in the air inside the cell.

The advantages of the non-dispersive infrared (NDIR) CO₂ sensors are selective, sensitive, non contact and reliable. At present time this technology is accepted as a state-of-the-art. But it still has some serious problems such as inherently expensive (at least two components), large size, requiring drift compensation and complex packaging.

So many efforts have been made regarding more miniaturization and lower system costs compared to the non-dispersive infrared CO₂ sensors.

POSIFA Microsystems Company announces a MEMS thermal conductivity CO₂ sensor.

The sensor integrates all CO₂ sensing components in a single silicon microstructure with a micro-hot-bridge, which is not like the non-dispersive infrared CO₂ sensor which is assembled with at least 4 separated components. In additional the micro-hot-bridge the sensor further contains a resistor, a cavity, and a thermopile with its hot junctions near the resistor and cold junctions extending to the silicon frame along the bridge supporting beams.  

The thermal conductivity CO₂ sensor performs a measurement as follows. By applying a voltage to the resistor on the micro-hot-bridge of the sensor, the resistor is heated up and becomes a “hot source”. The cavity is open to the atmosphere and filled with air containing a certain amount of CO₂ gas or air mixture. The air mixture transfers a quantity of heat from the hot source to the cold bottom of the cavity via the air mixture. The quantity of heat is measured by the thermopile. The changes in the thermal conductivity of the air mixture can be detected by measuring the changes of the output Seebeck voltage of the thermopile. With the measured thermal conductivity of the air mixture the concentration of the CO₂ in the air can be calculated by a humidity compensation algorithm which is based on the measurement results using the same sensor operated at two different temperatures.

Compared with the non-dispersive infrared CO₂ sensors, the thermal conductivity CO₂ sensors possess many advantages such as:

• Size reduced to a single silicon chip,
• Milliwatts grade power consumption,
• Milliseconds grade response times,
• Low cost, and
• Able to identify different gases of a gas mixture.

 

It is not be surprised for the last advantage. The thermal conductivities of gases always change with temperature. As shown in the above second figure, the thermal conductivities of gases CH₄, C₂H₆, N₂ and CO₂ increase slight-non-linearly with temperature. Based on this inherent character of the gases distinguishing single components of a gas mixture can be realized by modulating operation temperature of a thermal conductivity sensor.

Hear, as a gas mixture of Air, CO₂ and humidity (water vapor) is measured as a gas mixture. The temperature modulation is conducted by applying heating voltages V₁ and V₂ with V₂ higher than V₁. The measured output signals of the thermopile are Y₁ and Y₂ respectively. Then two binary linear equations can be obtained as

Y₁ = b₀₁ + b₁₁ N_CO2 + b₁₂ N_{water vapor}       (1)

Y₂ = b₀₂ + b₂₁ N_CO2 + b₂₂ N_{water vapor}        (2)

Where Y₁ and Y₂ are the output of the thermopile, N_CO2 and N_{water vapor}  are the volume percentage of CO₂ and humidity respectively, b_01, b_02, b₁₁, b_12, b_02, b₂₁ and b₂₂ are constants determined by calibration tests. The volume N_air of Air is found by meeting the equation as

N_air + N_CO2 + N_{water vapor} = 1                   (3)

With the measured Y_1, Y₂ and the known b_01, b_02, b₁₁, b_12, b_02, b₂₁ and b₂₂, N_CO2, and N_water can be found by solving the equations (1), (2) and (3).

All these advantages offer a high potential for mass markets. The most compact thermal conductivity CO₂ sensors have been developed for human breath analysis that will focus on enabling low cost applications but without compromising on accuracy.

Advantages of Thermal Conductivity Water Flow Sensors

over Plastic Spinning Water Flow Sensors

Xiang Zheng Tu

 

Reference to the above figure, a thermal conductivity water flow sensor is made up of two thermopiles, which is used as the sensing temperature difference element and operated in conjunction with a resistive heater element for thermoelectric sensing. The fabrication of such sensors is more complicated since less conventional materials are utilized for fabrication of thermopiles but CMOS (complementary metal oxide semiconductor) compatible processing is possible. The Seebeck effect of thermopiles enables higher sensitivity and unbiased output voltages with no offset or drift.

The thermopiles are constructed with thermocouples in series and so the output voltages due to temperature difference change is summed and increased over that of a single thermocouple. Since the thermal conduction between hot and cold junctions of the thermopiles and Johnson noise increases with increasing number of thermocouples, a high thermal isolation structure is desired in order to maximize temperature difference between hot and cold junctions.

The water mass flow (m) passing through the thermal conductivity sensor is calculated on the basis of the measured temperature difference (T_hot - T_cold) between the hot and cold junctions of the thermopile, and the thermal conductivity (C_p) coefficient (k), electric heat rate (q), and specific heat (C_p) of water, as follows:

m = kq/(C_p(T_hot – T_cold)                                    (1)

The electromotive force, or emf (V) created by the thermopile is directly proportional to the differential temperature (T_hot - T_cold) between the two junctions as

Emf_AB = nS_AB (T_hot - T_cold)                                (2)

Where n is the number of thermocouples of a thermopile and S (V/K) is called the Seebeck coefficient.

Still reference to the above figure, a plastic spinning water flow sensor has a rotor, a bearing, and a shaft, which are mounted in housing. The rotor spins as water passes over it. The measured flow rate is proportional to the rotational speed of the rotor. A variety of methods are used to detect the rotor speed, including a mechanical shaft and an electronic sensor.

Plastic bearings must be lubricated, not only to reduce friction and wear, but, in the case of plain bearings, to prevent them from seizing the shaft which they support. Self-lubricating plastic bearings contain a mix of dry lubricants. In operation, movement between shaft and bearing causes microscopic abrasion of the dry lubricant, filling and smoothing the shaft surface to reduce friction. The resulted micron particles will enter the water flow which is harmful to human health.

Most plastic bearing materials expand when exposed to heat and moisture. This factor is more significant when the running clearance between the bearing and shaft is less than 0.001 in. Plastic bearings and shafts are fabricated by injection molding process which has typical accuracy within 0.005 in. As a result, excessive wear or seizing of the shaft occurs very often.

Bubbles inevitably form as air is entrained in the water during the pouring process. The formed bubbles can create many problems in plastic spinning water flow sensors, such as:

● decreasing lubricity caused by an air emulsion,
● reduction of thermal conductivity,
● higher noise emission, and
● decrease water output efficiency.

Compare with plastic spinning water flow sensors, the thermal conductivity water flow sensors have the advantages as:

1.     Thermal conductivity water flow sensors have no moving parts, in which there are no any mechanical failures to take place.

2.     Thermal conductivity water flow sensors are MEMS devices with small size, higher sensitivity, higher reliability, low power consumption, ease of fabrication, and low cost.

3.     Thermal conductivity water flow sensors calculate mass flow rather than volumetric flow and do not require temperature or pressure correction, which means there is no additional expense for the purchase and installation of additional equipment.

4.     Thermal conductivity water flow sensors provide excellent accuracy and repeatability over a wide range of flow rates using bypass flow tube design. The sensor is placed in a bypass around a restriction in the main pipe and is sized to operate in the laminar flow region over its full operating range.

It should be emphasized that the thermal conductivity water flow sensors are not influenced by the air bubbles entrained in the water. The effect of the bubbles can be added to the series conductivity by using conductivity of the air-water mixture for the water conductivity. The thermal conductivity of continuous water phase with entrapped air bubbles can be calculated using Maxwell’s model as

k_m= k_c (k_d + 2k_c – 2p_d [ k_c –k_d])/(k_d + 2kc + p_d [k_c –k_d])            (3)

Where:

K_m = conductivity of the mixture,

K_c, k_d = conductivity of continuous and disperse phases, respectively, and

P_d = volume fraction of the disperse phase.

Replacing equation (3) into equation (1), the water mass flow rate measured by the thermal conductivity water flow sensor should be

 m = {(k_d + 2k_c – 2p_d [ k_c –k_d])/(k_d + 2kc + p_d [k_c –k_d])}q/(C_p(T_hot – T_cold)     (4)                        

It can be seen that the measured water mass flow rate does not contain the air bubbles entrained in the water.