Implementing Demons and Ratchets

Article Implementing Demons and Ratchets

Peter M. Orem * and Frank M. Orem ThermaWatts, LLC, Renton, WA 98057, USA; [email protected] * Correspondence: [email protected]; Tel.: +1-503-635-2607

Academic Editor: Kevin H. Knuth Received: 30 November 2016; Accepted: 6 January 2017; Published: 14 January 2017

Abstract: Experimental results show that ratchets may be implemented in semiconductor and chemical systems, bypassing the second law and opening up huge gains in energy production. This paper summarizes or describes experiments and results on systems that effect demons and ratchets operating in chemical or electrical domains. One creates temperature differences that can be harvested by a heat engine. A second produces light with only heat input. A third produces harvestable electrical potential directly. These systems share creating particles in one location, destroying them in another and moving them between locations by diffusion (Brownian motion). All absorb ambient heat as they produce other energy forms. None requires an external hot and cold side. The economic and social impacts of these conversions of ambient heat to work are, of course, well-understood and huge. The experimental results beg for serious work on the chance that they are valid.

Keywords: second law; thermodynamics; direct conversion; epicatalysis; semiconductor; LED; Brownian ratchet; Maxwell’s demon; ambient heat; energy

1. Introduction This paper is intended to assist understanding possible second law exceptions, so that readers might be moved to examine them carefully. The second law of thermodynamics was first expressed more than 190 years ago to show that there is an upper limit to the efficiency of a heat engine, which converts heat to mechanical energy or other work. Simply put, such conversion always suffers inefficiency. The second law has been thoroughly tested, restated and confirmed many times since then [1]. Scientists have shifted away from the notion of “laws” to ideas about “models” to explain and predict behaviors, because laws imply completeness, whereas models acknowledge their incompleteness. This gives freedom to expand our range of thinking, including that on the second law. A small set of researchers has made the transition to models as they develop experiments that point to new energy possibilities. They seem to be driven by the potentially huge value to humanity for finding such possibilities [2]. We present below three cases of systems that may be demonstrating demons and ratchets. These were chosen because:

1. They are supported by test results; 2. We believe they are conceptually clean and easily explained; and 3. They may have commercial application.

We believe that if any one of these demonstrations is valid (we believe all are), it opens up a range of practical methods for turning ambient heat directly into electrical energy. This matters because such direct conversion can be perfectly sustainable, since it does not require fuels or otherwise cause environmental damage.

Entropy 2017, 19, 34; doi:10.3390/e19010034 www.mdpi.com/journal/entropy Entropy 2017, 19, 34 2 of 19 Entropy 2017, 19, 34 2 of 19 The cases are: • The Epicatalysis: cases are: Tests demonstrate that opposing catalysts can create a temperature differential needed for a heat engine heat • Epicatalysis: Tests demonstrate that opposing catalysts can create a temperature differential • LEDs: Emission measurements show that unbiased light emitting diodes (LED’s) emit a needed for a heat engine heat. • LEDs:small Emission amount measurements of light. show that unbiased light emitting diodes (LED’s) emit a small •amount Doped of light.semiconductor structure: A sandwich-like structure of semiconductors presents a • Dopedpersistent, semiconductor harvestable structure: voltage. We A sandwich-likecall this an ambient structure thermal of semiconductorselectric converter presents (ATEC). a persistent, harvestable voltage. We call this an ambient thermal electric converter (ATEC). The demonstration systems have two features in common: particles are created and destroyed, and Thethese demonstration actions take place systems at different have two locations. features in common: particles are created and destroyed, and theseFigure actions 1 shows take the place general at different concept locations. for a cycle that would create an exception to the second law. Z representsFigure1 shows some potential the general or conceptwork function. for a cycle It co thatuld would be gravity, create electric an exception potential to the or secondsome other law. Zfunction. represents N represents some potential the number or work of function.particles in It the could system. be gravity, electric potential or some other function.The flow N represents is as follows: the number of particles in the system. The ﬂow is as follows: • The bottom line represents the creation of particles at Location 1. • •The Some bottom particles line represents move across the creation the potential of particles to Lo atcation Location 2, driven 1. by thermal energy through • Some particles move across the potential to Location 2, driven by thermal energy through diffusion. diffusion. This creation process fills the role of Maxwell’s demon. This creation process ﬁlls the role of Maxwell’s demon. • Some of those particles are then destroyed • Some of those particles are then destroyed. • • RemainingRemaining particles particles return return to Location to Location 1. 1.

FigureFigure 1.1. AA cyclecycle thatthat wouldwould createcreate aa secondsecond lawlaw exception.exception.

TheThe changingchanging numbernumber ofof particlesparticles ﬁllsfills thethe rolerole ofof aa BrownianBrownian ratchet.ratchet. ThereThere isis muchmuch ﬂexibilityflexibility inin implementingimplementing thisthis cycle,cycle, asas describeddescribed below.below. The The directional directional bias bias is is essential. essential. TheThe overalloverall cyclecycle convertsconverts thermalthermal energyenergy intointo somesome otherother formform ofof energyenergy andand thenthen preventsprevents thethe reversereverse transitiontransition by by removing removing the the particles particles involved. involved. Cheating Cheating the laws the oflaws physics of physics once is notonce enough; is not weenough; have towe do have it twice: to do bothit twice: Maxwell’s both Maxwell’s demon and demon a Brownian and a Brownian ratchet are ratchet required. are required. TheThe generalgeneral conceptconcept thatthat capturescaptures thethe objectionobjection to the cyclecycle of FigureFigure1 1 is is “detailed “detailed balance”. balance”. TheThe principleprinciple of of detailed detailed balance balance says says that that both both paths paths between between 1 and 1 and 2 should 2 should be in be equilibrium. in equilibrium. In that In regard,that regard, the principle the principle of detailed of detailed balance balance is the mostis the narrowly most narrowly focused focused reason whyreason such why cycles such should cycles notshould work. not In work. other In words, other thewords, principle the principle of detailed of detailed balance balance is a corollary is a corollary to the second to the law.second law. Note:Note: this paper only summarizes the methodsmethods andand resultsresults ofof epicatalysis,epicatalysis, partpart ofof thethe LEDLED emissionsemissions experimentsexperiments andand thethe dopeddoped semiconductorsemiconductor structure (ATEC) toto permitpermit thethe readerreader toto seesee thethe similaritiessimilarities ofof thethe systemssystems withoutwithout gettinggetting lostlost inin thethe details.details. Thus,Thus, thethe readerreader isis encouragedencouraged toto examineexamine thethe SupplementalSupplemental MaterialsMaterials andand referencesreferences forfor furtherfurther depthdepth of of understanding. understanding.

2. Materials and Methods

Entropy 2017, 19, 34 3 of 19

2.Entropy MaterialsEntropy 2017 2017, 19, and19, 34, 34 Methods 3 3of of 19 19

TheTheThe subsectionssubsections subsections belowbelow below describe describe the thethe setup setupsetup and and processes processes we we we used used used to to totest test test the the the hypotheses hypotheses hypotheses that that that certaincertaincertain systems systems systems convert convert convert heat heat heat to to workto work work without without without a colda acold cold side. side. side. We We intendWe intend intend that that thethat the information the information information in this in in thissection this andsectionsection our “A andand Report ourour on“A“A Testing ReportReport of onon ThermaWatts TestingTesting ofof PrototypeThermaWattsThermaWatts Ambient PrototypePrototype Thermal AmbientAmbient Electric ThermalThermal Converter ElectricElectric Devices” document,ConverterConverter accessibleDevices” Devices” document, throughdocument, the accessible accessible Supplementary through through Materials, the the Supplementary Supplementary are sufﬁcient Materials, Materials, for replication are are sufficient sufficient of our for workfor byreplicationreplication others. of of our our work work by by others. others.

2.1.2.1.2.1. LEDs LEDs LEDs TwoTwoTwo methods methods methods are are are describeddescribed described that that we we used used to toto captur capturecapture ethe the effect effect of of unbiased unbiased unbiased LEDs LEDs LEDs emitting emitting emitting light. light. light.

2.1.1.2.1.1.2.1.1. Direct Direct Direct Film Film Film ExposureExposure Exposure InInIn 2007–2008, 2007–2008, 2007–2008, Peter Peter Peter Orem Orem Orem set set set up up aup simplea asimple simple apparatus apparatus apparatus to tryto to try to try record to to record record photons photons photons that that mightthat might might be emitted be be fromemittedemitted an unbiased from from an an unbiased LEDunbiased upon LED LED photo-recombination upon upon phot photo-recombinationo-recombination of electrons of of electrons electrons and holes. and and holes. holes. TheTheThe apparatus apparatus apparatus is is asis as shownas shown shown in in Figurein Figure Figure2. Several2. 2. Several Several LEDs LEDs LEDs of various of of various various types types types were were mountedwere mounted mounted on a on printedon a a printed circuit board, each with its contact leads shorted together. High speed (Fuji Superia 3200, circuitprinted board, circuit each board, with each its contact with its leads contact shorted lead together.s shorted Hightogether. speed High (Fuji speed Superia (Fuji 3200, Superia Tokyo, 3200, Japan) Tokyo, Japan) color film was mounted in parallel to the board with LEDs. Working in a dark room, colorTokyo, film Japan) was mountedcolor film inwas parallel mounted to thein parallel board withto the LEDs. board Workingwith LEDs. in aWorking dark room, in a dark that assemblyroom, thatthat assembly assembly was was put put together together in in a alight-tight light-tight plastic plastic box box (Hammond (Hammond 1594EBK, 1594EBK, Guelph, Guelph, ON, ON, was put together in a light-tight plastic box (Hammond 1594EBK, Guelph, ON, Canada), which in turn Canada),Canada), which which in in turn turn was was sealed sealed with with black black tape tape. .The The assembly assembly sat sat on on a ashelf shelf for for several several months, months, was sealed with black tape. The assembly sat on a shelf for several months, after which the film was afterafter which which the the film film was was extracted extracted (again (again in in a adark dark room). room). The The film film was was then then taken taken to to a acommercial commercial extracted (again in a dark room). The film was then taken to a commercial service for development. serviceservice for for development. development.

FigureFigureFigure 2. 2. Apparatus Apparatus for forfor recording recordingrecording ph ph photonsotonsotons from from from unbiased unbiased unbiased LEDs. LEDs. LEDs.

SampleSampleSample exposures exposures exposures areare are shownshown shown in in Section Section 33. .3.

2.1.2.2.1.2.2.1.2. Digital Digital Digital ImagingImaging Imaging ofof of Heated, Heated, Unbiased Unbiased LED LEDLED LikeLikeLike the the the direct direct direct ﬁlm film film exposure exposure exposure above,above, above, thethe the apparatusappa apparatusratus waswas was devised devised to to capture capture photons photons that thatthat might mightmight be emittedbebe emitted emitted from from anfrom unbiased an an unbiased unbiased LED uponLED LED upon recombinationupon recombin recombination ofation electrons of of electrons electrons and holes. and and holes. Theholes. apparatus The The apparatus apparatus is shown is is in Figureshownshown3. in in Figure Figure 3. 3.

FigureFigure 3. 3. ApparatusApparatus Apparatus for for recording recording from from from heated heated heated LED. LED. LED.

Entropy 20172017,, 1919,, 3434 4 of 19

The camera is a ProEM 512® from Princeton Instruments (Trenton, NJ, USA). It uses their lowest ® pixelThe resolution camera isto a minimize ProEM 512 thefrom exposure Princeton time Instruments needed for (Trenton, each pixel. NJ, USA).It has It a uses relatively their lowest high pixelquantum resolution efficiency to minimize to capture the near exposure infrared time from needed thefor LED. each A pixel.light Itfilter has is a relativelymounted highin the quantum lens. A efficiencycomputer towith capture Princeton near infraredInstrument’s from LightSource the LED. A® light software filter Version is mounted 6.0 drives in the lens.camera A computeroperation ® withand captures Princeton image Instrument’s output. LightSourceThe LED is heldsoftware against Version a heat 6.0 sink/distributor drives camera (for operation good heat and capturestransfer) imageby a spring output. clip. The Twisted LED is heldconductors against are a heat run sink/distributor from the LED leads (for goodto a switchable heat transfer) bias by source a spring (open, clip. Twistedshorted,conductors +1.5 V, 1.5 areV). runThe from heat thesink/distributor LED leads to ahas switchable heating biasresistors source attached (open,shorted, (2 × 20 Ω +1.5 in V,parallel, 1.5 V). The50 W), heat which sink/distributor are in turn driven has heating at up resistorsto 90 watts attached from (2a variable× 20 Ω inDC parallel, power 50supply. W), which Temperature are in turn at driventhe heat at sink/distributor up to 90 watts fromis measured a variable by DC a thermocouple. power supply. The Temperature heat sink/distributor, at the heat sink/distributorheater and LED isassembly measured is housed by a thermocouple. in a fiberglass-filled The heat wood sink/distributor, box and covered heater with and aluminum LED assembly foil. The is housedcamera inand a fiberglass-filledbox are mounted wood firmly box together and covered to maintain with aluminum focal distance foil. to The the camera LED. Tests and boxare arerun mountedin a dark firmlyroom. togetherThe tolight maintain filter and focal target distance LED to are the selected LED. Tests to have are runmatching in a dark wavelength room. s at 940 nm. The light filterThe is a lightThorLabs filter bandpass and target filter, LED CWL are selected = 940, FWHM to have = matching 10 (centered wavelengths at 940 nm, at with 940 a nm. width The of light ±10 filternm). The is a ThorLabsLED is a Fairchild bandpass semiconductor filter, CWL = 940,QEB363ZR—LED FWHM = 10 (centeredIR EMITTING at 940 940 nm, NM with 2MM a width ZB T/R of ±(Digikey10 nm). QEB363ZRCT-ND, The LED is a Fairchild Phoenix, semiconductor AZ, USA). QEB363ZR—LED IR EMITTING 940 NM 2MM ZB T/R (DigikeyThe tests QEB363ZRCT-ND, may be run with Phoenix, the AZ, heat USA). sink/distributor held at a temperature (per thermocouple/voltmeter)The tests may be run in with steady the heatstate sink/distributorat up to 196 °C. Exposure held at a temperaturetimes are typically (per thermocouple/ up to an hour ◦ voltmeter)long per frame. in steady state at up to 196 C. Exposure times are typically up to an hour long per frame. Sample exposures areare shownshown inin thethe Results.Results. 2.2. Doped Semiconductor Structures (ATEC) 2.2. Doped Semiconductor Structures (ATEC) An ATEC is a semiconductor sandwich of high and low density semiconductors, with alternate An ATEC is a semiconductor sandwich of high and low density semiconductors, with alternate n-doped and p-doped layers between, as described in Section 4.2. n-doped and p-doped layers between, as described in Section 4.2. 2.2.1. The ATEC Structure 2.2.1. The ATEC Structure Several prototype ATECs were built at Stanford University per the specifications listed in the Several prototype ATECs were built at Stanford University per the specifications listed in the Supplemental Materials, Test Results, Section2: Conceptual Design. Supplemental Materials, Test Results, Section 2: Conceptual Design. Figure4 shows one of the ATEC prototypes we had fabricated. Figure 4 shows one of the ATEC prototypes we had fabricated. 1 • It is 10 ⁄2 cycles of AlGaAs with either 33% or 50% Al, 5.5 µm thick • It is 10 ½ cycles of AlGaAs with either 33% or 50% Al, 5.5 µm thick • It is on a base of n-GaAs, 0.5 mm thick and 75 mm (3”) in diameter • It is on a base of n-GaAs, 0.5 mm thick and 75 mm (3”) in diameter • It has a gold conductor final layer. The surface features you see are reflections of the room. • It has a gold conductor final layer. The surface features you see are reflections of the room.

Figure 4. Prototype ambient thermal electric converter (ATEC).

2.2.2. Simulation Tools Modeling used Silvaco’s Atlas 5.16.3.R, which includes simulation software and semiconductor materials’ data.

Entropy 2017, 19, 34 5 of 19 Entropy 2017, 19, 34 5 of 19 2.2.3. Laboratory Test Tools

2.2.3.The Laboratory custom test Test jig Tools and tools were developed to address the considerable challenges of testing (see the Supplemental Materials, Test Results, and Section 3: Model Results). Primary challenges are listedThe below. custom test jig and tools were developed to address the considerable challenges of testing (see the Supplemental Materials, Test Results, and Section3: Model Results). Primary challenges are listed• below.Costs: since our prototype ATEC material set was constrained by costs, we could look for only microvolt outputs, which are hard to measure without high-end equipment (also • Costs: since our prototype ATEC material set was constrained by costs, we could look for only cost-constrained) microvolt outputs, which are hard to measure without high-end equipment (also cost-constrained) • • Seebeck Seebeck effect effect • • Voltage Voltage across across the the ATEC ATEC from from electrical electrical noise no inducedise induced through through wiring wiring and and the ATECthe ATEC and thenand rectiﬁedthen rectified by the ATEC, by the which ATEC, is which inherently is inherently a diode a diode • •Meter Meter bias bias

WeWe are looking for any output voltage across the ATECATEC atat zerozero input.input. ToTo establish confidence conﬁdence and credibility for our test method, we collected data points either side ofof zero input: a a sweep. The The sweep sweep order order was was prog programmedrammed to to run run from from a a small positive voltage to maximummaximum positive positive voltage voltage and and from from small small negative negative voltage voltage down down to to maximum maximum negative negative voltage. voltage. There was a 40-s delay from sweep voltage setting to voltage measurement mitigating the the effects of systemsystem capacitance. capacitance. The The zero zero sweep sweep voltage voltage point point ou outputtput should should fall fall on on the the straight line of outputs fromfrom the the rest rest of of the the sweep. sweep. Interim Interim data data were were record recordeded and and inspected inspected for for stability. stability. All sweep sweep data data are are availableavailable at the ThermaWatts, LLC website [[3].3]. FigureFigure 55 isis thethe top-leveltop-level viewview ofof thethe testtest apparatus.apparatus. ItIt showsshows thethe sourcesource ofof sweepsweep voltagevoltage andand thethe measurementmeasurement of net voltage (sweep voltage plus any contribution by the ATEC) acrossacross thethe ATEC.ATEC.

FigureFigure 5. 5. ATECATEC with with sweeping voltage source.

OurOur approach approach to tackle the several challenges was to: • • ApplyApply a known a known sweep sweep voltage voltage to theto the ATEC ATEC and and a known a known resistor resistor in series in series • • Amplify Amplify the the voltage voltage across across the the known known resistor resistor to to bring bring it it from from thethe microvoltmicrovolt to millivolt range range • •Put Put the the test test target, target, measurement measurement circuitry circuitry and and wiring wiring (as (as muchmuch asas possible)possible) intointo the oven (environmental(environmental chamber) chamber) • •Design Design all lowall low voltage voltage signals signals such such that that they they are inare a in small a small region region at a uniformat a uniform temperature, temperature, with short wires; longer runs were ampliﬁed (or not yet divided) by 1000 or 4000, respectively. with short wires; longer runs were amplified (or not yet divided) by 1000 or 4000, Figurerespectively.6 shows a somewhat more detailed view of the test jig used to meet these challenges while ﬁnding the zero sweep voltage output of the ATEC. Figure 6 shows a somewhat more detailed view of the test jig used to meet these challenges The complete circuit for driving the sweep voltage and detecting net voltage is shown in Figure7. while finding the zero sweep voltage output of the ATEC. It is described in the Supplemental Materials, Test Results, and Section5: Test Setups and Procedures. The complete circuit for driving the sweep voltage and detecting net voltage is shown in Figure 7. It is described in the Supplemental Materials, Test Results, and Section 5: Test Setups and Procedures.

Entropy 2017, 19, 34 6 of 19 EntropyEntropy 2017 2017, 19, 19, 34, 34 6 6of of 19 19

Entropy 2017, 19, 34 6 of 19

Figure 6. Test jig. FigureFigure 6. TestTest jig. jig.

Figure 6. Test jig.

FigureFigureFigure 7. 7. Test Test jigjig jig circuitry. circuitry.circuitry.

Figure 8 shows: FigureFigure8 shows: 8 shows: A.A. ATECATEC on on a astand, stand, with with wiring, wiring,Figure inside inside 7. Testan an environmental jigenvironmental circuitry. chamber. chamber. A. ATECB.B. onATEC aATEC stand, on on a with acustom custom wiring, ceramic ceramic inside stand, stand, an with environmentalwith contact contact wires wires chamber. Figure 8 shows: B. ATECC. on aRelay custom control ceramic board stand,and Arduino with contact processor wires. for sweep control logic. C.A. RelayATEC control on a stand, board with and wiring, Arduino inside processor an environmental for sweep control chamber. logic. C. Relay control board and Arduino processor for sweep control logic. B. ATEC on a custom ceramic stand, with contact wires C. Relay control board and Arduino processor for sweep control logic.

(A:(A: chamber chamber) ) (B:(B: stand stand) ) (C:(C: relay relay board board) ) Figure 8. Sample test jig apparatus. Figure 8. Sample test jig apparatus. (A: chamber) (B: stand) (C: relay board) Figure 8. Sample test jig apparatus. Figure 8. Sample test jig apparatus.

Entropy 2017, 19, 34 7 of 19 Entropy 2017, 19, 34 7 of 19

3. Results The results of the two experimentalexperimental areas identiﬁedidentified in the Materials and Methods are summarized here:

1. 1.UnbiasedUnbiased LEDs LEDs emitting emitting light light 2. 2.Harvestable Harvestable voltage voltage across across a doped a doped semiconductor semiconductor structure structure (ATEC) (ATEC) Additional data are available in the Supplemental Materials, Test Results, and Section 4: Test Additional data are available in the Supplemental Materials, Test Results, and Section4: Results. Test Results.

3.1. LEDs Test resultsresults toto determinedetermine if if unbiased unbiased LEDs LEDs emit emit light light consist consist of of photographic photographic images images collected collected by differentby different methods methods (direct (direct ﬁlm film exposure exposure and and digital digital images). images).

3.1.1. Direct Film Exposure In our firstfirst experiment on LED emissions, aa strip of filmfilm was exposed to a set of LEDs (each with its leadsleads tied tied together) together) for for a period a period of several of several months months (see Section (see Section 2.1.1 in the2.1.1 Materials in the Materials and Methods). and WhenMethods). developed, When developed, the film showed the film frames showed like theframes one inlike Figure the one9 (which in Figure contains 9 (which the clearest contains effect the weclearest were effect looking we for,were while looking containing for, while no containing obvious flaws). no obvious The glowing flaws). areas,The glowing one small areas, and one focused, small oneand largefocused, and one blurred, large areand arguably blurred, theare resultarguably of photons the result emitted of photons from twoemitted LEDs. from two LEDs.

Figure 9. Film exposed to shorted LED.

3.1.2. Digital Imaging of Heated, Unbiased LED Images from our second experiment for LED emissions,emissions, set up per 2.1.2 in the MaterialsMaterials and Methods, are shown and described below.below. Figure 10 shows thethe cameracamera viewview ofof aa 940-nm940-nm LEDLED mountedmounted onon thethe heatheat sink/distributorsink/distributor within the digitaldigital imagingimaging box. Frame Frame A A was taken by a digitaldigital camera in roomroom lighting.lighting. The LED is the golden circle, upper left. The black material is rosinrosin from soldering. Frames B and C show the view as seen by the Princeton InstrumentsInstruments ProEMProEM 512512®® camera through LightFieldLightField®® 6.0 software, with the heat sink/distributor warmed warmed to to 196 196 °C.◦C. The The test test r roomoom was was dark. dark. Exposure Exposure time time for for each each was was 30 30 min. min. The onlyonly knownknown differences differences between between the the images images in Framesin Frames B and B and C is C the is connectionthe connection of the of LED the leads.LED Forleads. Frame For Frame B, the leadsB, the areleads open. are Foropen. Frame For Frame C, the C, leads the areleads connected are connected to battery to battery that applies that applies a 1.5-V a reverse1.5-V reverse bias. bias. TheFigure images 10 shows in Frames the camera B and view C are of distinctlya 940-nm differentLED mounted in the on region the heat of the sink/distributor LED (above thewithin red arrow)the digital and imaging otherwise box. quite Frame similar. A was In Frame taken C,by there a digital appears camera to be in less room light lighting. at the LED. The LED is the golden circle, upper left. The black material is rosin from soldering. Frames B and C show the view as seen by the Princeton Instruments ProEM 512® camera through LightField® 6.0 software, with the heat sink/distributor warmed to 196 °C. The test room was dark. Exposure time for each was 30 min. The only known differences between the images in Frames B and C is the connection of the LED leads. For Frame B, the leads are open. For Frame C, the leads are connected to battery that applies a 1.5-V reverse bias.

Entropy 20172017,, 1919,, 3434 88 ofof 1919

EntropyThe2017 images, 19, 34 in Frames B and C are distinctly different in the region of the LED (above the8 ofred 19 arrow) and otherwise quite similar. In Frame C, there appears to be less light at the LED.

((A: visible light)) ((B: IR shorted)()(C: IR reverse bias))

FigureFigure 10. 10. ImagesImagesImages fromfrom from experimentexperiment experiment onon on digitaldigital digital imagingimaging imaging ofof of heatedheated heated LED.LED. LED. RedRed Red ArrowsArrows Arrows touchtouch touch rectangularrectangular rectangular basebase of of LED, LED, lens lens is is above arrow tip.

3.2. Doped Semiconductor Structure 3.2. Doped Doped Semiconductor Structure Testing of our ATEC consisted of modeling the structure and then running extensive Testing of our ATEC consistedconsisted of modeling the structure and then running extensive measurements against prototype ATECs. measurements against prototype ATECs.

3.2.1.3.2.1. Simulation Simulation BeforeBefore we we had the prototype ATEC fabricated,fabricated, wewe ranran modelsmodels onon itsits structurestructure andandand materials.materials.materials. FigureFigure 1111 showsshows modeled modeled behavior behavior of ofof AlGaAs AlGaAsAlGaAs in theinin the conﬁgurationthe configurconfiguration for an for ATEC. an ATEC. The graph The showsgraph showsa non-zero a non-zero ﬂow of electronsflow of electrons and holes and through holes the th ATEC.roughrough Additionalthethe ATEC.ATEC. resultsAdditionalAdditional are in resultsresults the Supplemental areare inin thethe SupplementalMaterials, Test Materials, Results, and Test Section Result3s,: Modeland Section Results. 3: Model Results. TheseThese results results encouraged us to move to to prototyping prototyping and and laboratory tests.tests.

Figure 11. Modeled flow of electrons and holes. Figure 11. ModeledModeled flow ﬂow of electrons and holes.

3.2.2. Laboratory Tests 3.2.2. Laboratory Laboratory Tests Prototype ATECs were tested, with considerable care given to holding sources of error to Prototype ATECsATECs were were tested, tested, with with considerable considerable care givencare given to holding to holding sources ofsources error toof negligible error to negligible levels. As an imposed voltage was swept across a prototype, the measured output voltage negligiblelevels. As levels. an imposed As an imposed voltage voltage was swept was acrossswept aacross prototype, a prototype, the measured the measured output output voltage voltage was was consistently different from the imposed voltage (see the Supplemental Materials, Test Results, wasconsistently consistently different different from from the imposedthe imposed voltage voltag (seee (see the Supplementalthe Supplemental Materials, Materials, Test Test Results, Results, and and Section 4: Test Results). andSection Section4: Test 4: Results).Test Results).

Entropy 2017, 19, 34 9 of 19 Entropy 2017, 19, 34 9 of 19

Figure 12 shows a plot of the current on the y-axis vs. sweep voltage on the x-axis, at 120 ◦F. Figure 12 shows a plot of the current on the y-axis vs. sweep voltage on the x-axis, at 120 °F. The graph shows: The graph shows:

• •Parallel Parallel lines lines across across the the sweep sweep of bias,of bias, one one for fo eachr each orientation orientation of theof the ATEC. ATEC. The The blue blue dots dots are from the “normal” ATEC orientation, and the red dots are for “reverse.” are from the “normal” ATEC orientation, and the red dots are for “reverse.” • Linear offset of voltage across the target. • Linear offset of voltage across the target.

Figure 12. Current/Voltage Plot for ATEC R2 Prototype at 120 °F. Figure 12. Current/Voltage Plot for ATEC R2 Prototype at 120 ◦F.

The plots for voltage across the ATEC do not pass through zero current at zero voltage. The offsetsThe between plots for those voltage plots across are opposite the ATEC and do about not pass equal through at zero zero bias. current at zero voltage. The offsets betweenThis thosestrongly plots suggests are opposite that the and prototype about equal contri atbuted zero bias. its own voltage (half the distance between the lines),This strongly either with suggests or in that opposition the prototype to the contributed imposed itsvoltage, own voltage depending (half theon distancewhich way between the prototypethe lines), eitherwas turned. with or in opposition to the imposed voltage, depending on which way the prototype was turned.We take the difference as twice the contribution to voltage made by the target ATEC FigureWe take 13 the shows difference the results as twice from the runs contribution made at 160 to voltage °F with made an ATEC by the and target then ATEC with resistors of ◦ 160 KFigure Ohm and13 shows 220 K theOhm. results from runs made at 160 F with an ATEC and then with resistors of 160 K Ohm and 220 K Ohm. • The ATEC normal and reverse run results are again separated and again do not pass • Thethrough ATEC normal zero current and reverse at zero run voltage; results are again separated and again do not pass through zero current at zero voltage; • The resistor plots both pass through zero (at this scale) and have different slopes, as expected. • The resistor plots both pass through zero (at this scale) and have different slopes, as expected. We also ran tests for photosensitivity, flooding the target with AC noise and the effects of flexingWe the also target ran tests (see for the photosensitivity, Supplemental Materials, ﬂooding the Test target Results, with ACand noiseSection and 4: the Test effects Results). of ﬂexing The resultsthe target were (see consistent the Supplemental with the models. Materials, Test Results, and Section4: Test Results). The results were consistentWe conclude with the that models. the target ATEC is producing electricity and surmise that it is absorbing heat to satisfyWe conclude the conservation that the targetof energy. ATEC is producing electricity and surmise that it is absorbing heat to satisfy the conservation of energy.

Entropy 2017, 19, 34 10 of 19 Entropy 2017, 19, 34 10 of 19

Figure 13. CurrentCurrent vs. Voltage Voltage Plot for ATEC R2 Prototype andand resistorsresistors atat 160160 ◦°F.F.

4. Discussion This section section describes describes systems systems under under test test that that demonstrate demonstrate conversion conversion of ambient of ambient heat to heat useful to usefulwork. work.

4.1. Epicatalysis Catalysts havehave beenbeen usedused toto assistassist inin chemicalchemical reactions reactions for for about about 200 200 years. years. They They may may speed speed up up a reactiona reaction and and lower lower energy energy requirements, requirements, but do not bu changet do thenot chemical change equilibriumthe chemical (completeness equilibrium of a(completeness reaction). While of a they reaction). participate While in they reactions, participate they in are reactions, not consumed. they are not consumed. The term “epicatalysis” is used to identify pairs of catalysts that oppose each other in assisting a chemical reaction,reaction, such such that that their their chemical chemical equilibrium equilibrium points points are different. are different. The idea The has idea been has around been foraround more for than more 100 than years, 100 with years, several with supportingseveral supporting experiments. experiments. If oneone puts puts two two catalysts catalysts with with opposing opposing characteristics characteristics in contact in contact with a bathwith ofa chemicalbath of reactants,chemical perreactants, Figure per 14, Figure three things 14, three will things occur: will occur: • Catalyst 1 one will create Reactant Set A in favor of Reactant Set B, and Catalyst 2 will create • Catalyst 1 one will create Reactant Set A in favor of Reactant Set B, and Catalyst 2 will create Reactant Set B in favor of Reactant Set A. Reactant Set B in favor of Reactant Set A. • • By diffusion,By diffusion, the the bath bath will will tend tend to to equalize equalize the the reactants reactants in in the the reverse reverse direction,direction, BB diffusing towardtoward Catalyst Catalyst 1 and 1 and A toward A toward Catalyst Catalyst 2. 2. • •One One of the of reactionsthe reactions (A => (A B => or B B or => B A) => will A) will generate generate more more net heat net (beheat exothermic) (be exothermic) than thethan other, the so theother, catalyst so the surface catalyst will warm,surface and will the warm, other willand absorbthe other more will net absorb heat (be more endothermic). net heat (be

Theendothermic). difference in net heat (and therefore temperature) may be used to drive a heat engine with no externalThe colddifference side. in net heat (and therefore temperature) may be used to drive a heat engine with no external cold side.

Entropy 2017, 19, 34 11 of 19 EntropyEntropy 2017 2017, ,19 19, ,34 34 1111 of of 19 19

Figure 14. A view of epicatalysis. FigureFigure 14. 14. A A view view of of epicatalysis. epicatalysis. Since the reactants and the catalysts are not consumed, the process will continue indefinitely. If heat SinceisSince added the the toreactants reactants the bath, and and the the the temperature catalysts catalysts are are are differenc not not not consumed, consumed, consumed,e at the the catalyststhe the process process process may will will willbe continue continueharvested, continue indefinitely. indefinitely. indeﬁnitely.perhaps as If If electricityIfheatheat heat isis is addedadded added using toto tothe thethe the Seebeck bath,bath, bath, thethe theeffect, temperaturetemperature temperature indefinitely. differencdifferenc difference ee atat the the catalysts catalysts may maymay bebe harvested, harvested, perhapsperhaps asas electricityelectricityDr. Daniel usingusing the Sheehanthe SeebeckSeebeck Seebeck [4–6] effect,effect, effect, and indeﬁnitely.indefinitely. indefinitely. David Miller [7] have experimental data on temperature differencesDr.Dr. Daniel DanielDaniel created Sheehan SheehanSheehan by epicatalysis. [4 – [4–6]6[4–6]] and and Davidand Such DavidDavid Miller differenc MillerMiller [7] havees [should[77] experimental] havehave support experimentalexperimental data a heat on temperatureengine datadata on onoperating temperaturetemperature differences only withcreateddifferencesdifferences ambient by epicatalysis. createdcreated heat (no byby cold epicatalysis. Suchepicatalysis. side). differences SuchSuch should differencdifferenc supporteses shouldshould a heat supportsupport engine a operatinga heatheat engineengine only operatingoperating with ambient onlyonly heatwithwithSheehan (noambient ambient cold side).heat usedheat (no (nohydrogen cold cold side). side). as the bath and catalysts of rhenium (Re) and tungsten (W). The Sheehan used hydrogen as the bath and catalysts of rhenium (Re) and tungsten (W). The reactionsSheehan were usedused molecular hydrogen hydrogen hydrogen as as the the bath H 2bath to and atomic and catalysts catalystshydrogen of rhenium of H rheniumat the (Re) rhenium and (Re) tungsten and catalyst tungsten (W). and The H (W). reactions to H The2 at reactions were molecular hydrogen H2 to atomic hydrogen H at the rhenium catalyst and H to H2 at thewerereactions tungsten molecular were catalyst, molecular hydrogen as shown Hhydrogen2 to atomicin Figure H2 hydrogen to 15. atomic See H[6] hydrogen at for the complete rhenium H at theinterpretation catalyst rhenium and Hcatalyst of to Figure H2 atand the 15. H tungsten to H2 at catalyst,thethe tungsten tungsten as shown catalyst, catalyst, in Figure as as shown shown 15. Seein in Figure Figure [6] for 15. 15. complete Se Seee [6] [6] interpretationforfor complete complete interpretation interpretation of Figure 15. of of Figure Figure 15. 15.

Figure 15. Setup for epicatalysis tests. (a) Overview of test bed heating and temperature measurement;FigureFigure 15. 15.15.Setup SetupSetup (b) for Detail for epicatalysisfor epicatalysiswithepicatalysis catalytic tests. tests.( tests.asurfaces) Overview ((aa) )inside OverviewOverview of testchamber. bed ofof heating testHavetest bed bedgot and theheating temperatureheating permission andand measurement; temperatureoftemperature Springer. (measurement;bmeasurement;) Detail with ( catalytic(bb)) Detail Detail surfaces with with catalytic catalytic inside surfaces chamber.surfaces inside inside Have chamber. chamber. got the permission Have Have got got the ofthe Springer. permission permission of of Springer. Springer. Because these reactions only take place significantly at high temperatures, the bath and Because these reactions only take place significantly at high temperatures, the bath and measurementBecause these apparatusthese reactions reactions were only heatedonly take placetake to 1950 significantlyplace °K. signifSheehan aticantly high measured temperatures,at high the temperatures, difference the bath in and temperaturethe measurement bath and at measurementmeasurement apparatus apparatus were were heated heated to to 1950 1950 °K. °K. Sh Sheehaneehan measured measured the the difference difference in in temperature temperature at at theapparatus catalysts were at heatedup to to126 1950 °K ◦(seeK. Sheehan Figure measured16, which the shows difference the temperature in temperature difference at the catalysts between at thethe catalystscatalysts atat upup toto 126126 °K°K (see(see FigureFigure 16,16, whichwhich showsshows thethe temperaturetemperature differencedifference betweenbetween catalystsup to 126 ◦asK (seetemperature Figure 16 ,increases, which shows and the see temperature [6] for complete difference interpretation between catalysts of Figure as temperature 16). This catalystscatalysts asas temperaturetemperature increases,increases, andand seesee [6][6] foforr completecomplete interpretationinterpretation ofof FigureFigure 16).16). ThisThis temperatureincreases, and difference see [6] for could complete be harvested interpretation as electr of Figureicity indefinitely, 16). This temperature with heat supplied difference as could part beof temperaturetemperature differencedifference couldcould bebe harvestedharvested asas electrelectricityicity indefinitely,indefinitely, withwith heatheat suppliedsupplied asas partpart ofof maintainingharvested as bath electricity temperature. indefinitely, with heat supplied as part of maintaining bath temperature. maintainingmaintaining bath bath temperature. temperature.

FigureFigure 16. 16. TemperatureTemperature differences between catalysts [[66]. Have got the permission of Springer. FigureFigure 16. 16. Temperature Temperature differences differences between between catalysts catalysts [6 [6].]. Have Have got got the the permission permission of of Springer. Springer.

Entropy 2017, 19, 34 12 of 19

David Miller [7] is working toward parallel results using formic acid and other low-strength hydrogen bond chemicals with catalyst surfaces, such as Teﬂon, and at typical environmental temperatures. He has achieved a temperature differential of about 0.2 ◦C. His work continues to explore combinations of materials and catalysis surfaces. These two demonstrations are elegant and conclusive and alone should be sufﬁcient to excite the energy world.

4.2. LED Emissions at Zero Bias LEDs with a zero bias may continually emit photons, as shown in our and others’ experiments and for the reasons explained below.

4.2.1. Direct Film Exposure Photon emissions of unbiased LEDs seem a plausible explanation for the glow spots captured by direct ﬁlm exposure. There are, however, alternate explanations to consider.

• Photographic film is actually tricky stuff. It is sort of a chemical bath. It reacts to photons, but the chemical images therefrom degrade over time, smearing and fading. It may be that the long exposure time used actually led to reduced image capture. • Some of the frames had large crescent images, which did not seem to correspond to LEDs. These may have resulted from bending the film during mounting or dismounting, so we discounted those images. • The “light-tight box” may not have been so. We discount this because of the box assembly and because the shape and location of images are inconsistent with how light might have leaked in. • The film may have interacted chemically with the LED lens. We discount that because such an interaction should have occurred with many of the LEDs, which seems not to be reflected in the images. • The LED tails might be acting as antenna to electrical noise, providing an oscillating pump across the LEDs. Although the tails are small and twisted together, we cannot discount this possibility. Were we to rerun this experiment, we would want to put the box in a Faraday cage (perhaps a wrap of metal foil or mesh); or we could perhaps measure the antenna effect of similarly-sized wires across a diode.

In balance, although the experiment was informal and incomplete, the results were sufﬁcient to encourage further work.

4.2.2. Over-Unity LED Output A team at MIT [8,9] set up precise measurements of electrical power input to and light output from an LED in a temperature-controlled environment. The test consisted of putting a carefully-measured amount of power (as electricity) into the LED and measuring its light output. The key results for tests made at three temperatures are shown in Figure 17. The circled point in Figure 17, at very nearly zero input, shows the light output exceeding the power input and rising as input neared zero. We know of no reason why there would be a discontinuity in light output between the last-measured bias and zero bias. Lacking such discontinuity, a reasonable explanation for the experimental results is that emissions continue through (and past) the zero bias point. The MIT results tend to corroborate our own testing of unbiased LED for photon emissions. Entropy 2017, 19, 34 13 of 19 Entropy 2017, 19, 34 13 of 19

Figure 17. Over-unity efficiency of LED near zero bias [10]. Figure 17. Over-unity efﬁciency of LED near zero bias [10].

4.2.3. Digital Imaging of Heated, Unbiased LED 4.2.3. Digital Imaging of Heated, Unbiased LED Largely because the direct film exposure test was lengthy and therefore hard to reproduce, we Largely because the direct film exposure test was lengthy and therefore hard to reproduce, we developed an equivalent test for LED emissions using a high-sensitivity camera. Even with such a developed an equivalent test for LED emissions using a high-sensitivity camera. Even with such a camera, it is necessary to heat an LED to bring its light output up to the camera’s range for practical camera, it is necessary to heat an LED to bring its light output up to the camera’s range for practical exposure times. exposure times. The LED light output consists of blackbody emissions resulting from various emissivities of The LED light output consists of blackbody emissions resulting from various emissivities of materials in the target area and, arguably, emissions from recombination photon emissions. The materials in the target area and, arguably, emissions from recombination photon emissions. The output output curves for these as a function of temperature are similar, making the separation of effects curves for these as a function of temperature are similar, making the separation of effects difficult to difficult to detect. Our solution was to: detect. Our solution was to: • Choose an LED with a wavelength where its output is strongest at a high temperature • Choose an LED with a wavelength where its output is strongest at a high temperature (therefore, some(therefore, red). some red) • • Balance Balance that that wavelength wavelength against against the quantum the quantum efficiency efficiency of the camera, of the socamera, that drop-off so that would drop-off not be toowould severe. not be too severe. • •Filter Filter the the output output of the of the LED LED (blackbody (blackbody and and possible possible recombination recombination emission) emission) to the to camerathe camera at a wavelength matching the LED. at a wavelength matching the LED. • Observe the image captured for open and closed connection to the LED leads and with reverse • Observe the image captured for open and closed connection to the LED leads and with bias on the leads. Sufficient reverse bias will surely block electron leaps across the LED junction. If therereverse are bias photons on the being leads. generated Sufficient at noreverse bias, thebias image will surely with reverseblock electron bias should leaps be across different the fromLED the junction. other two. If there are photons being generated at no bias, the image with reverse bias should be different from the other two. The images shown in Section 3.1.2 show such a difference. The comparison of Frames B and C of FigureThe 10 imagesshows thatshown we reducedin Section the 3.1.2 light show by blocking such a difference. LED operation The withcomparison a reverse of bias. Frames We concludeB and C ofthat Figure the blocked 10 shows light that was we from reduced the LED the emitting light by light blocking without LED bias, operation as was towith be demonstrated.a reverse bias. We concludeThere that are sourcesthe blocked of error light and was misinterpretation from the LED in theseemitting test light results: without bias, as was to be demonstrated. • The temperatures of test runs with and without reverse bias could differ. We discount this because Thereconsiderable are sources care wasof error taken and to controlmisinterpretation temperature, in these and settling test results: time was allowed. This might be improved further with automated control of the heater. • The temperatures of test runs with and without reverse bias could differ. We discount this • The exposure time could vary between test runs. We discount this because exposure times were softwarebecause controlled. considerable care was taken to control temperature, and settling time was allowed. • TheThis setup might does be not improved completely further control with the automated location of control its image of the in theheater. field. While true, the LED •shape The is exposure pretty clearly time identifiable,could vary between so this appears test runs. not We to be discount a problem. this because exposure times • Noisewere could software be creating controlled. the apparent differences in images. We discount this because noise •tends The to setup be uniformly does not distributedcompletely acrosscontrol the the field, locati whereason of its the image area in of the the field. LED imageWhile istrue, clearly the consistently locatable. LED shape is pretty clearly identifiable, so this appears not to be a problem.

Entropy 2017, 19, 34 14 of 19

• Noise could be creating the apparent differences in images. We discount this because noise tends to be uniformly distributed across the field, whereas the area of the LED image is Entropy 2017clearly, 19, 34 consistently locatable. 14 of 19

The images can be improved to decrease noise, either by: The images can be improved to decrease noise, either by: • Post-processing the images with any of several techniques • Post-processing the images with any of several techniques • Using a yet-more-sensitive camera, perhaps a liquid nitrogen-cooled camera, such as • Using a yet-more-sensitive camera, perhaps a liquid nitrogen-cooled camera, such as Princeton Instrument’sPrinceton PylonInstrument’s® CCD BRXPylon (Trenton,® CCD BRX NJ, USA).(Trenton, NJ, USA).

4.2.4.4.2.4. RationaleRationale forfor LightLight EmissionsEmissions fromfrom UnbiasedUnbiased LEDsLEDs LightLight emissionemission atat thethe zero-biasedzero-biased LEDLED arisesarises fromfrom thethe randomrandom behaviorbehavior ofof electronselectrons withinwithin theirtheir semiconductorsemiconductor materials.materials. Figure 18 shows a generally-acceptedgenerally-accepted distributiondistribution ofof electronelectron energyenergy inin semiconductors.semiconductors. It says that some ofof thethe electronselectrons havehave higherhigher energyenergy thanthan average.average. CarriersCarriers areare createdcreated atat thethe metalmetal contacts,contacts, andand somesome ofof themthem acquireacquire enoughenough energyenergy fromfrom thethe latticelattice toto reachreach thethe directdirect recombinationrecombination regions regions of of the the LED, LED, where where they they release release some some of their of their energy energy as a photon.as a photon. The LED The mustLED coolmust to cool reﬂect to thisreflect energy this loss,energy but loss, is rewarmed but is rewarmed from itsenvironment. from its environment. Thus, heat Thus, is converted heat is toconverted light. to light.

FigureFigure 18.18. DistributionDistribution ofof electronelectron energyenergy (Fermi–Dirac).(Fermi–Dirac).

4.3.4.3. DopedDoped SemiconductorSemiconductor StructureStructure (ATEC)(ATEC) OurOur measurementsmeasurements forfor voltagevoltage acrossacross anan ATECATEC show show consistently consistently (absolute) (absolute) positive positive results. results. OperationOperation ofof an an ATEC ATEC starts starts with “churning”with “churning” of carriers of (dynamic carriers equilibrium)(dynamic equilibrium) at a heterojunction at a ofheterojunction semiconductors of semiconductors of differing intrinsic of differing carrier intrinsic densities. carrier Churning densities. is our Churning term for is what our mustterm for be happeningwhat must atbe the happening heterojunction. at the heterojunction. We have not measured We have itnot directly measured (only it indirectly directly (only through indirectly ATEC performance).through ATEC Physics performance). does not talkPhysics about does it much not (theretalk about has been it much no apparent (there reasonhas been to). no Yet, apparent it has to bereason happening to). Yet, because: it has to be happening because:

• •Generation Generation in a in semiconductor a semiconductor material material at a at given a given temperature temperature is a functionis a function of the of bandgap.the bandgap. • •The The recombination recombination rate rate in in a a semiconductorsemiconductor material material is is a function a function of carrier of carrier density density and band and band structure. structure. • Generation and recombination produce an intrinsic (equilibrium) carrier density, which is a • Generation and recombination produce an intrinsic (equilibrium) carrier density, which is a function of the bandgap and band structure. • Somefunction materials of the have bandga bandp and structures band structure. that support higher carrier densities than others, •independent Some materials of the bandgap.have band structures that support higher carrier densities than others, • Carriersindependent move by of diffusion the bandgap. from higher density to lower density regions (given no electrical force on them), including across a heterojunction. • If two materials with different band structures are joined at a heterojunction, the transport across the junction will typically be out of equilibrium with the internal generation/recombination equilibrium of each material. Entropy 2017, 19, 34 15 of 19 Entropy 2017, 19, 34 15 of 19 • Carriers move by diffusion from higher density to lower density regions (given no electrical • forceCarriers on them),move by including diffusion across from ahigher heterojunction. density to lower density regions (given no electrical • Ifforce two on materials them), including with different across band a heterojunction. structures are joined at a heterojunction, the transport Entropy• 2017, 19, 34 15 of 19 acrossIf two materialsthe junction with different will typicallyband structures be out are joinedof equilibrium at a heterojunction, with thethe transportinternal generation/recombinaacross the junctiontion will equilibrium typically of eachbe material.out of equilibrium with the internal • Each material will move to balance at its intrinsic density (for a given temperature) through a • Eachgeneration/recombina material will movetion to equilibrium balance at its of intrinsiceach material. density (for a given temperature) through change in rate of recombination. • aEach change material in rate will of moverecombination. to balance at its intrinsic density (for a given temperature) through Wea have change a numerical in rate of modelrecombination. for the ﬂow of carriers in churning, thanks to Yang et al., 1993 [11]. We have a numerical model for the flow of carriers in churning, thanks to Yang et al., 1993 [11]. They only identify the transport model, not the full process with generation and recombination. TheyWe only have identify a numerical the transport model model, for the not flow the of full carriers process in churning, with generation thanks andto Yang recombination. et al., 1993 [11]. Graphically, churning might look like this: Figure 19 is a simple representation of semiconductors TheyGraphically, only identify churningthe transport might model, look not likethe fullthis: process Figure with 19 generation is a simple and recombination.representation of with differing intrinsic densities, not joined. semiconductorsGraphically, with churning differing might intrinsic look densities, like this: not joined.Figure 19 is a simple representation of semiconductors with differing intrinsic densities, not joined.

Figure 19. Semiconductors of different intrinsic carrier densities. Figure 19. Semiconductors of different intrinsic carrier densities. Figure 20 depicts the results of diffusion as th thee differing semiconductor materials materials are are joined. joined. Immediately,Figure 20 some depicts of thethe carriersresultscarriers of diffusediffuse diffusion toto thethe as otherthothere differing material.material. semiconductor It is exaggerated materials for explanatory are joined. purposes,Immediately, showing some oneof the half carriers of carriers diffuse moving to theacross. other The material. point is Itthat is densit densityexaggeratedy differences for explanatory (resulting frompurposes, band showing structure) one drive half the of carriers netnet ﬂow.flow. moving across. The point is that density differences (resulting from band structure) drive the net flow.

Figure 20. Diffusion across a heterojunction. Figure 20. Diffusion across a heterojunction. As diffusion occurs, carrier densities change. Generation in each semiconductor continues, essentiallyAs diffusion unaffected occurs,occurs, by density carriercarrier densitiesdensitieschange. Recombin change.change. Generationation Generation drops in inin each the highersemiconductor density material continues, as carriersessentially diffuse unaffectedunaffected away by byand density density rises change. change.in the Recombination lowerRecombin densityation drops materialdrops in the in higher asthe carriers higher density densityarrive. material Thus,material as carriers each as semiconductordiffusecarriers awaydiffuse and responds away rises inand to the restorerises lower in it density theto itslower intrinsic material density density. as carriers material Diffusion arrive. as carriers continues. Thus, eacharrive. Generation semiconductor Thus, eachand recombinationrespondssemiconductor to restore continues.responds it to itsWeto intrinsicrestore call this it density. continuousto its intrinsic Diffusion cycle density. “churning”. continues. Diffusion Generation continues. and Generation recombination and continues.recombinationWe can We calculate callcontinues. this continuousthe We net call flow this cycle ofcontinuous “churning”.carriers fromcycle high“churning”. carrier density to low carrier density materialsWe can canthat calculate calculate we used the thein net ATECnet ﬂow flow ofprototypes, carriers of carriers from AlGaAs from high carrierwithhigh 0.5carrier density and density0.33 to low Al, carrier torespectively, low density carrier based materials density on Yangthatmaterials we et usedal. that [11]. in we ATECThe used following prototypes, in ATEC equations AlGaAsprototypes, from with AlGaAsthe 0.5 paper and with 0.33 for Al, 0.5electron respectively,and 0.33and Al,hole based respectively, motion on Yang are evaluated etbased al. [ 11on]. TheinYang Table following et al.1. [11]. equations The following from the equations paper for from electron the paper and holefor electron motion areand evaluated hole motion in Table are evaluated1. in Table 1. −+ΔE Jqvn=−(1 +δ ) (0− ) × exp( −Δ∆EC ) + qvn (1 +δ ) (0+ ) Jn,i ni=,1−qvn n(121 + δ)n1(0 −+) × exp(− E ) + qv nn (1 + δ)n 22(0 ) Jqvn=−1 (1 +δ ) (0 ) × exp( −kTkTC ) + qvn2 (1 +δ ) (0 ) ni,1 n12ΔkT n 2 −+Ev Jqvp=−(0− )exp( −∆ΔEv ) − qvp 0+ J pi,1= −qv n12p (0 −+) exp(− E ) − qv p p 20 Jqvpp,i =−n1 1 (0 )exp( −kT v ) − qvpp2 2 0 pi,1 n12kT p 2 Material values are from Silvaco’s ATLAS, exceptexcept forfor afﬁnity,affinity, wherewhere we we used used the the 60/40 60/40 rule.rule. Material values are from Silvaco’s ATLAS, except for affinity, where we used the 60/40 rule.

Entropy 2017, 19, 34 16 of 19

Table 1. Values for calculating currents across AlGaAs heterojunction.

Al Fractions Units Notes and References 0.33 0.5 - - ni 3.82 × 103 1.52× 103 /cm3 Carrier density at 310 K, per ATLAS Entropy 2017, 19, 34 16 of 19 Eg 1.78 1.97 0.19 eV Per ATLAS

∆Ec - 0.6 0.114 eV 60/40 rule for AlGaAs Table 1. Values for calculating currents across AlGaAs heterojunction. ∆Ev - 0.4 0.076 eV 60/40 rule for AlGaAs

kBT - Al Fractions - 0.0267138 eV Units kB from Notes ATLAS and References

∆Ec/kBT - 0.33 - 0.5 4.27 --- - ni 3.82 × 103 1.52× 103 /cm3 Carrier density at 310 K, per ATLAS ∆Ev/kBT - - 2.84 - - Eg 1.78 1.97 0.19 eV Per ATLAS −2 exp(∆E∆c Ec/kBT) - - - 0.6 1.40 0.114 × 10 - eV - 60/40 rule for AlGaAs ∆Ev - 0.4 0.076 eV 60/40 rule for AlGaAs −2 exp(∆Ev/kBT) - - 5.81 × 10 - - kBT - - 0.0267138 eV kB from ATLAS vn∆ Ec/kBT3.86 - × 107 1.34 - × 108 - 4.27 cm/s - Yang - [11] Equations (5) and (9) ∆Ev/kBT - - 2.84 - - 7 8 −2 vpexp( ∆Ec/kBT) 1.42-- × 10 1.40 × 10 - 1.40 × 10 cm/s-- Yang [11] Equations (12) and (13) exp(∆E /k T) -- 5.81 × 10−2 -- Densitiesv B - - - - - vn 3.86 × 107 1.34 × 108 - cm/s Yang [11] Equations (5) and (9) n vp 6.181.42 ×× 1010 7 3.901.40 ×× 1010 8 - - /cm cm/s3 - Yang [11] Equations (12) and (13) Densities ----- 3 p n 6.18× 6.18 ×1010 3.90 3.90 ×× 1010 - - /cm/cm3 - - Currentsp - 6.18× 10 - 3.90 × 10Net - -/cm 3 - - Currents -- Net -- 7 8 8 2 n 7 8 8 2 J Jn −−3.343.34 ×× 1010 5.225.22 ×× 1010 4.894.89 ×× 1010 /cm/cm/s/s NetNet of 0.33 of 0.33 => => 0.50 0.50 minus minus 0.50 0.50 => 0.33 0.33 for electronselectrons 7 8 8 2 J 5.10 × 107 −5.46 × 108 −4.95 × 108 /cm2 /s Net of 0.33 => 0.50 minus 0.50 => 0.33 for holes Jp p 5.10 × 10 −5.46 × 10 −4.95 × 10 /cm /s Net of 0.33 => 0.50 minus 0.50 => 0.33 for holes

We can relate thethe currentcurrent ofof anan ATECATEC prototypeprototype withwith thethe currentcurrent computedcomputed above.above. NetNet JJnn andand JJpp are nearly equalequal inin countcount at at about about 4.9 4.9× ×10 107 7carriers carriers per per cm cm2 s2 s or or about about 7.9 7.9× ×10 10−11−11 A/cmA/cm2.. Because Because the charges are opposite, this works out to roughly zero net current. The prototype area is about 35 cm2,, and the the current current was was about about 2.3 2.3 × 10× −1011 A− 11or A6.5 or × 6.510−12× A/cm10−122. TheA/cm lower2. The prototype lower prototypecurrent is accounted current is foraccounted in part forby inthe part effect by of the doped effect oflayers doped and layers the li andmited the thickness limited thickness of x = 0.50 of xlayers. = 0.50 Figure layers. 21 Figure below 21 relatesbelow relatesthese quantities. these quantities.

Figure 21. AlGaAs Heterojunction Bands and Currents.

For churning, we further note that: • • While While there there is a is flow a flow of carriers, of carriers, there there is no is netno net flow flow of charge. of charge. • There must be a flow of heat in the opposite direction of carriers, since generation is endothermic

and recombination is exothermic. Because this creates a temperature differential, it is in itself an exception to the second law. The temperature difference measurement was cited by George Levy [12], which references offsets encountered by Iwanaga et al. [13]. • This is a generally useless and sometimes annoying feature of heterojunctions (so far). Entropy 2017, 19, 34 17 of 19

• There must be a flow of heat in the opposite direction of carriers, since generation is endothermic and recombination is exothermic. Because this creates a temperature differential,

Entropy 2017it ,is19 in, 34 itself an exception to the second law. The temperature difference measurement17 was of 19 cited by George Levy [12], which references offsets encountered by Iwanaga et al. [13]. • This is a generally useless and sometimes annoying feature of heterojunctions (so far). Here is another way of looking at churning, as though a temperature phenomenon. Let us imagineHere that is another the carrier way density of looking difference at churning, was the resultas though of a temperaturea temperature difference, phenomenon. rather Let than us a imagineband structure that the difference. carrier density Effectively, difference a material was the with result a higher of a temperature C acts as if it difference, is hotter than rather a material than a bandwith astructure lower C. difference. From the intrinsic Effectively, carrier a material density: with a higher C acts as if it is hotter than a material with a lower C. From the intrinsic carrier density: 3 n*p = C*T *exp(−Eg/kB*T) n*p = C*T3*exp(−Eg/kB*T) Let us suppose thethe currentscurrents arearethe the result result of of different different carrier carrier densities densities in in the the same same material material due due to todifferent different temperatures. temperatures. What What temperature temperatur woulde would produce produce the the same same current current ratio? ratio?

3 3 3 3 (n*p)/(n(n*p)/(n00*p0)) = = (J (Jnn*J*Jpp)/(Jn)/(Jn0*Jp0*Jp0) =0) C =0*T C0*T*exp(*exp(−Eg0−/kEBg0*T)/C/kB*T)/C0*T0 *exp(0*T0−E*exp(g0/kB*T−E0)g0 /kB*T0) where T0 = 310 K, Eg is 1.78 eV (for x = 0.33) and the currents are taken from Table 1. Solving where T0 = 310 K, Eg is 1.78 eV (for x = 0.33) and the currents are taken from Table1. Solving numerically, numerically,we get a temperature we get a oftemperature 334.5 K. of 334.5 K. Given a temperature difference, we can use the S Seebeckeebeck effect across interposed doped layers to produce aa voltage,voltage, as as shown shown in Figurein Figure 22. This22. isThis done is usingdone dopingusing todoping control to the control Seebeck the coefficient: Seebeck coefficient:p-doping giving p-doping a positive giving coefficient, a positive such coefficient, that holes such flow that from hole “hot”s flow to “cold”, from “hot” and n-doping, to “cold”, giving and n-doping,the reverse giving effect the (similar reverse to aeffect hot point (similar probe). to a hot point probe).

Figure 22. Seebeck effect across doped layers. Figure 22. Seebeck effect across doped layers.

In the AlGaAs sample, most of the ratio is taken up by the exponent term. For x = 0.33 at 310 K, the exponentIn the AlGaAs is −66, so sample, a small mostpercentage of the change ratio is in taken the exponent up by the has exponenta large change term. inFor the xvalue. = 0.33 Forat − our310 K,HgCdTe the exponent models, is the66, soexponent a small percentageis less than change −10, which in the exponentresults in has much a large larger change effective in the − temperaturevalue. For our differences. HgCdTe models, the exponent is less than 10, which results in much larger effective temperatureWe made differences. a semiconductor sandwich of high and low density semiconductors, with alternate n-dopedWe madeand p-doped a semiconductor layers between, sandwich as ofshown high in and Figure low density 23. The semiconductors, carriers try to churn, with alternate but the dopingn-doped forces and p-doped electrons layers only between,to the left as and shown holes in only Figure to 23the. Theright, carriers as shown try to in churn, Figure but 23a. the After doping an initialforces surge electrons of churning, only to the a balance left and is holes reached only (like to thethat right, which as creates shown a in built-in Figure potential 23a. After in an LED), initial butsurge there of churning, is no net aflow balance across is reachedlayers because (like that carriers which have creates nowhere a built-in to go. potential If the sandwich in an LED), has but a conductorthere is no netlayer ﬂow added across at layers each becauseend and carriers connected, have nowhereelectrons to may go. Ifflow the sandwichfrom the hasleft aconductor, conductor throughlayer added a load, at each to the end right and conductor, connected, as electrons shown in may Figure ﬂow 23b. from the left conductor, through a load, to the right conductor, as shown in Figure 23b.

Entropy 2017, 19, 34 18 of 19 Entropy 2017, 19, 34 18 of 19

(a) (b)

Figure 23. Churning with doped layers. ( aa)) High High and and low density layers with n-doped and p-doped layerslayers added; added; ( b) Complete assembly of layers with contacts and load added.

5. Conclusions 5. Conclusions We have presented experimental results from three systems, each of which indicates that those We have presented experimental results from three systems, each of which indicates that those systems draw heat from their environments and turn it into harvestable work. While the LED direct systems draw heat from their environments and turn it into harvestable work. While the LED direct film exposure experiment is admittedly light, the other experiments were serious, careful work. Each ﬁlm exposure experiment is admittedly light, the other experiments were serious, careful work. Each of of them contributes to the picture that we need to move from laws to relevant models: them contributes to the picture that we need to move from laws to relevant models: • Epicatalysis created a harvestable temperature difference. That, alone, is enough to prove • Epicatalysis created a harvestable temperature difference. That, alone, is enough to prove we we must move past laws. must move past laws. • • The The light light emitted emitted by unbiased by unbiased LEDs, LEDs, as captured as captur byed the by experiments, the experiments, could could in theory in theory be harvested. be Whileharvested. not as dramatic While not as as epicatalysis, dramatic as it epicatalysis, also proves theit also point. proves the point. • •Consistent Consistent voltage voltage measurements measurements from a a do dopedped semiconductor semiconductor structure structure also demand also demand lookinglooking forward. forward.

We believebelieve thatthat implementingimplementing demons demons and and ratchets ratchets is possible.is possible. We We believe believe the openthe open question question now nowis whether is whether it leads it toleads practical to practical results. results. There mayThere be may many be more many ways more to ways implement to implement them; Figure them;1 Figureshould 1 prove should a usefulprove toola useful for hunting tool for them.hunting them.

5.1. Interpretation Interpretation We see a huge practical potential for the ATEC. Let us look at some numbers to see where this might go.

• •For For AlGaAs, AlGaAs, the the largest largest bandgap bandgap that that we we used used in the in the ATEC ATEC construction construction is 1.79 is 1.79 eV. eV. • •The The test test results results for ourfor AlGaAsour AlGaAs prototypes prototypes are, we are, think, we think, signiﬁcant. significant. However, However, they do they not screamdo not “useful”. We remind you that we used AlGaAs because many shops could fabricate it. scream “useful”. We remind you that we used AlGaAs because many shops could fabricate it.

But consider HgCdTe (Mercat), for whichwhich wewe areare lookinglooking atat aa largestlargest bandgapbandgap ofof 0.210.21 eV.eV. 13 With Mercat, it should be possible to achieve about 1013 improvementimprovement in in output output or or kilowatts kilowatts per per cubic centimeter centimeter of of the the layered layered materi material.al. It Itis isenough enough that that the the limiting limiting factor factor is heat is heat transport transport into intothe lattice.the lattice. In terms In terms of power of power output, output, it does it does not not look look like like a abattery battery so so much much as as a thermally-charged capacitor, ableable toto releaserelease rapidly rapidly its its thermal thermal energy energy down down to ato low a low temperature. temperature. Production Production costs costs may mayreach reach as low as aslow pennies as pennies per watt per watt capacity. capacity.

5.2. Areas for Research and for Engineering 5.2. Areas for Research and for Engineering There is much work to do to bring to life the principles identiﬁed. We see: There is much work to do to bring to life the principles identified. We see:

• Replicating experiments in epicatalysis, LEDs and ATECs;

Entropy 2017, 19, 34 19 of 19

• Replicating experiments in epicatalysis, LEDs and ATECs; • Using better material sets for ATECs and lower temperature epicatalysis; • Developing application-speciﬁc packages at least for ATECs (largely an engineering process); • Looking for other instances of particle creation and destruction, coupled with opportunities for the diffusion of particles (demons and ratchets), for other useful exceptions to the second law.

Supplementary Materials: The following is available online at www.mdpi.com/1099-4300/9/1/34/s1. A Report on Testing of ThermaWatts Prototype Ambient Thermal Electric Converter Devices, V1.1, Peter Orem, 12/2013. Acknowledgments: Funding for this work came from Frank and Judy Orem and from Peter Orem. Princeton Instruments has graciously loaned cameras to us. We have received questions and ideas from the Oregon State University Physics Department. Author Contributions: Peter Orem was the major contributor to this work, from conception to test methods and testing. Frank Orem contributed suggestions on testing and was the principal writer. Conﬂicts of Interest: The authors declare no conﬂict of interest.

References

1. Nikulov, A.V.; Sheehan, D.P. Special Issue on Quantum Limits to the Second Law of Thermodynamics. Entropy 2004, 6, 1–10. [CrossRef] 2. Sheehan, D.P. Second law of thermodynamics: Status and challenges. In Proceedings of the AIP Conference, San Diego, CA, USA, 14–15 June 2011. 3. ThermaWatts, LLC. Available online: http://www.ThermaWatts.com (accessed on 11 January 2017). 4. Sheehan, D.P. Nonequilibrium heterogenous catalysis in long term mean-free-path regime. Phys. Rev. E 2013, 88, 032125. [CrossRef][PubMed] 5. Sheehan, D.P.; Garamella, J.T.; Mallin, D.J.; Sheehan, W.F. Steady-state non-equilibrium temperature gradients in hydrogen gas-metal systems: Challenging the second law of thermodynamics. Phys. Scr. 2012, 2012, 014030. [CrossRef] 6. Sheehan, D.P.; Mallin, D.J.; Garamella, J.T.; Sheehan, W.F. Experimental Test of a Thermodynamic Paradox. Found. Phys. 2014, 44, 235–247. [CrossRef] 7. Miller, D. Apparatus and Procedures for Detecting Temperature Differentials with Epicatalytic Materials. Entropy 2017, submitted. 8. Santhanam, P.; Gray, D.J.; Ram, R.J. Thermoelectrically pumped light emitting diodes operating above unity efﬁciency. Phys. Rev. Lett. 2012, 108, 097403. [CrossRef][PubMed] 9. Gray, D. Thermal Pumping of Light Emitting Diodes. Master’s Thesis, Massachusetts Institute of Technology, Cambridge, MA, USA, 2011. 10. Santhanam, P.; Dodd, J.G.; Ram, R.J. Thermo-Electrically Pumped Light-Emitting Diodes. U.S. Patent US 20140159582, 12 June 2014. 11. Yang, K.Y.; East, J.R.; Haddad, G.I. Numerical Modeling of Abrupt Heterojunctions Using a Thermionic-Field Emission Boundary Condition. Solid State Electron. 1993, 36, 321–330. [CrossRef] 12. Levy, G.S. Thermoelectric Effects under Adiabatic Conditions. Entropy 2013, 15, 4700–4715. [CrossRef] 13. Iwanaga, S.; Toberer, E.S.; LaLonde, A.; Snyder, G.J. A High Temperature Apparatus for Measurement of Seebeck Coefﬁcient. Rev. Sci. Instrum. 2011, 82, 063905. [CrossRef][PubMed]

© 2017 by the authors; licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC-BY) license (http://creativecommons.org/licenses/by/4.0/).