Farming thermoelectric paper

Bacteria are used to grow in an aqueous medium a cellulose-carbon nanotube composite porous film with good thermoelectric properties, flexibility and recyclability.

6 One Laser Raman Thermometry One laser Raman thermometry (1LRT) experiments were performed in backscattering geometry using a LabRam HR800 spectrometer with a grating of 1800 lines per millimeter. The setup is equipped with a liquid-nitrogen cooled charge coupled device (CCD) detector. A 488 nm line from an Ar + -ion gas laser was used as excitation source. A 50x long working distance Olympus objective (NA 0.50) was used to focus the laser light yielding an intensity Gaussian beam of 2.2 μm in diameter; the beam size was determined by scanning the spot across a cleaved silicon edge with a high resolution motorized stage. The thickness of the samples was measured with a digital micrometer with an accuracy of ± 0.5 μm. The free-standing films were held by a metallic ring and loaded onto a Linkam cryostat in ultrahigh vacuum (P < 10 -5 hPa) to suppress convection and potential sample degradation. The thermal contact of the samples with the Linkam hot stage was assured using silicon oil at the interfaces with the metal ring. Prior to starting the 1LRT experiments, the composite samples were baked at 120 ºC for 20-30 minutes to remove any water absorbed by the bacterial cellulose, which is known to be hygroscopic. The determination of the Raman shift coefficients was done right after the baking, while cooling back down from 120 to 30 ºC in steps of 10-20 ºC (Supplementary Figure 5a).
The power absorbed was determined by measuring the transmitted and reflected components for each incident power selected. The transmitted power was measured by coupling a power meter to the backside of the Linkam stage in a home-made setup that assured perfect sealing. The reflected component was measured by collecting the reflected laser intensity at the CCD and then using a mirror of known reflectivity as calibration guideline. Both the transmitted and incident powers also required calibration and correction for the reflection at the Linkam windows.
The heat equation was solved in COMSOL Multiphysics using cylindrical 2D symmetry in models of at least 2 mm lateral size. Radiation was included in the calculations, which required the experimental determination of the emissivity. For this, we used an Optris PI450 infrared camera finding a very close value for all CNT containing samples (ε = 0.94). The temperature raise was also weighted by the Gaussian shape of the heat source (i.e. the laser) to take into account the collection yield of the Raman-scattered light. While keeping the rest of model inputs fixed, the thermal conductivity was swept in the range of absorbed laser powers used in each experiment (typically 20-400 μW) leading to a linear increase of the temperature raise as a function of the absorbed power (Supplementary Figure 5b). Since the slope inversely depends on the thermal conductivity, the thermal conductivity of the samples was determined by fitting the results of the simulation to the experimentally found trend.   With P 0 the total dissipated power at the resistor and l the length of the resistor.
Since the AC current frequency determines the thermal penetration depth according to 1/ = √ / 2 , for thick films supported on a semi-infinite medium (i.e. a substrate) the Δ vs ln 2 curve shows low and mid-frequency regimes that primarily correspond to the substrate and the supported film, respectively. From the slope ( Δ / ln 2 ) of both regions it is straightforward to determine their thermal conductivity.
Statistics performed on 7 different resistors thermally evaporated on glass-supported BC films yielded an average thermal conductivity of (0.52 ± 0.05) W m -1 K -1 , which is in very good agreement with the theoretical values reported elsewhere.