writing lab reports - Are you sure you want to look at this?

pages.physics.cornell.edu
  • No tags were found...

writing lab reports - Are you sure you want to look at this?

writing lab reportsPhD Comics - Jorge Cham 09/14/11Physics 6510 / 4410Fall 2011Kyle Shenthanks to Don Hartill, Peter Wittich, others for materials


purpose of publication in science“if you haven’t published it, it didn’t happen”• Establishes precedence and ownership ofexperimental results• Communicates results to the scientificcommunity : colleagues (and competitors!)• Influences outcome of future proposalsand support for your research• Provides your patron (the public) with arecord of what you’re doing with their $$$


modeled on a Physical Review Letter (PRL)• 4 pages, single spaced, two columns(~ 2500 words)• prl.aps.org (look up your favorite topic anddownload a paper), or check references• standalone, self-contained document thatdescribes results completely• goal to get practice writing a real scientificpaper


what a PRL typically looks like (4 pages)PRL 99, 187001 (2007)PHYSICAL REVIEW LETTERS week ending2 NOVEMBER 2007PRL 99, 187001 (2007)PHYSICAL REVIEW LETTERS week ending2 NOVEMBER 2007Evolution of the Fermi Surface and Quasiparticle Renormalizationthrough a van Hove Singularity in Sr 2 y La y RuO 4K. M. Shen, 1,2, * N. Kikugawa, 3,† C. Bergemann, 4 L. Balicas, 5 F. Baumberger, 1,3 W. Meevasana, 1 N. J. C. Ingle, 1,2Y. Maeno, 6,7 Z.-X. Shen, 1 and A. P. Mackenzie 31 Departments of Applied Physics, Physics, and Stanford Synchrotron Radiation Laboratory, Stanford University,Stanford, California 94305, USA2 Department of Physics and Astronomy, University of British Columbia, Vancouver, B.C., V6T 1Z4, Canada3 School of Physics & Astronomy and Scottish Universities Physics Alliance, University of St. Andrews,North Haugh, St. Andrews KY16 9SS, United Kingdom4 Cavendish Laboratory, University of Cambridge, J. J. Thomson Avenue, Cambridge CB3 OHE, United Kingdom5 National High Magnetic Field Laboratory, Florida State University, Tallahassee Florida 32306, USA6 International Innovation Center, Kyoto University, Kyoto 606-8501, Japan7 Department of Physics, Graduate School of Science, Kyoto University, Kyoto 606-8502, Japan(Received 26 March 2007; published 29 October 2007)We employ a combination of chemical substitution and angle resolved photoemission spectroscopy toprove that the Fermi level in the band of Sr 2 y La y RuO 4 can be made to traverse a van Hove singularity.Remarkably, the large mass renormalization has little dependence on either k or doping. By combiningthe results from photoemission with thermodynamic measurements on the same batches of crystals, wededuce a parametrization of the full many-body quasiparticle dispersion in Sr 2 RuO 4 which extends fromthe Fermi level to approximately 20 meV above it.DOI: 10.1103/PhysRevLett.99.187001 PACS numbers: 74.20.Rp, 74.25.Jb, 74.72. h, 79.60. iThe layered perovskite ruthenate Sr 2 RuO 4 has attractedconsiderable attention and interest, stimulated initially byits unconventional superconductivity [1,2]. There is now aconsiderable body of evidence that it is a spin tripletsuperconductor with a time-reversal symmetry breakingground state [2–5] and that the metal from which thesuperconductivity condenses is a nearly two-dimensionalFermi liquid with a substantial mass enhancement [6]. Theelectronic structure near the Fermi level consists of twoFermi surface sheets ( and ) having strong Ru d xz;yzcharacter and a third, , based on the in-plane Ru d xyorbital.The comprehensive knowledge that has been acquiredabout the Fermi liquid properties of Sr 2 RuO 4 is unprecedentedfor a transition metal oxide, making it an idealmaterial in which to address a deeper layer of importantquestions, for example the origin of the mass enhancementand its relationship to the microscopic mechanism of thesuperconductivity. Recently, a new route to probing thiswas opened by the observation that the heterovalent substitutionof La 3 for Sr 2 to form Sr 2 y La y RuO 4 producesa rigid-band shift of the Fermi level for low y [7]. The limiton those experiments was the window for observing ade Haas–van Alphen (dHvA) signal, which is exponentiallysuppressed by disorder scattering, and not the solubilitylimit of La 3 in Sr 2 RuO 4 . In fact, much highersubstitution levels of up to y 0:27 have been achieved,raising the possibility that the Fermi level can be forced totraverse a van Hove singularity (vHS) in the band. Thisissue is of great interest, since strong electronic renormalizationand scattering of quasiparticles could occur nearsuch a singularity in the density of states. For instance, ithas even been postulated that some of the unusual physicalproperties observed in the cuprate superconductors mightbe as a result of the vicinity of a vHS to E F at k ; 0 .In this Letter we report the results of a careful set ofexperiments which prove conclusively that the mass renormalizationremains remarkably independent of both k anddoping as the vHS traverses E F .In principle, angle resolved photoemission spectroscopy(ARPES) is ideally suited to this experiment, since it is notsubject to the same purity restrictions as dHvA. However,since it is highly surface-sensitive there is always theconcern that some of the information that it yields is aproperty of the surface rather than the bulk electronicstructure. This issue has been of particular significance toARPES studies of Sr 2 RuO 4 , as discussed in Refs. [2,8,9].For this reason we set ourselves the more ambitious goal ofperforming a detailed comparison of ARPES with bulkmeasurements of the Fermi surface topography and excitationspectrum. The ARPES work was performed with aGammadata VUV5000 He-discharge lamp, monochromator,and a SES2002 electron analyzer, at a base temperatureof 10 K and a pressure of better than 4 10 11 torr.All data shown in this Letter were collected using He IIphotons (40.81 eV) on samples cleaved at 200 K, in orderto suppress the surface-related features discussed inRefs. [8,9]. A new generation of dHvA measurementswas also performed in the world’s highest steady magneticfield of 45 T using using a Cu-Be torque magnetometerwith capacitive readout, at temperatures between 0.04 and0.9 K.0031-9007=07=99(18)=187001(4) 187001-1 © 2007 The American Physical SocietyEnergy (meV)0-50-100a)0.8y = 010 K1.0also verify that these probes yield an electronic structureconsistent with one expected from measuring bulk thermodynamicproperties (specific heat). In Fig. 3(a) we presentthe direct comparison of Fermi surface topography measurementsby dHvA and ARPES across a range of cationsubstituted single crystals. By using the 45 T hybrid magnetat the U. S. National High Magnetic Field Laboratory inTallahassee, we were able to measure all three main dHvAfrequencies out to y 0:06, and the smallest frequencyall the way to y 0:1.Our findings have significant implications for the physicsof many-body renormalization in Sr 2 RuO 4 . If thisrenormalization is dominated by coupling to a sharpmode, it results in a kink in the dispersion at a characteristicenergy associated with that mode. If, on the otherhand, it is dominated by coupling to a continuum ofelectron-hole excitations, renormalization of the wholeband is possible. For this to occur, the source of therenormalization must be local in real space since its effectsare spread out over all k. A notable example is the simplifiedon-site repulsion U employed to partially accountfor correlations in modern electronic structure calculations.The extent of renormalization can also be linked tothe magnitude of the bare density of states, so naively onemight expect a feedback effect in which the correlationinducedenhancement factor is increased near the vHS.Several features of our data suggest that, perhaps surprisingly,the renormalization of the band is dominated by aglobal band renormalization. The data from across therange of La dopings that we have employed fall onto asingle cosine, even on the most crucial -M- cut throughthe vHS, and has a total bandwidth approximately a factorof 6 smaller than the LDA prediction, in agreement, withinexperimental error, with the dHvA-measured mass renormalizationat y 0 of m =m band 5:5 [2]. It can also betracked over a wide range of k (at least 20% of theBrillouin zone dimension), confirming that the source ofthe renormalization is highly local in real space.1.2b)0.8y = 0100 K1.01.2c)0.8y = 0.1810 K1.01.2k x(π/a)The data in Fig. 2(e) are from a restricted cut in theBrillouin zone, and do not, in isolation, allow us to drawfirm conclusions about the renormalized band shape for theremainder of the band or that of the and bands.Indeed, attempting to do this using ARPES would be rathercomplex [15]. However, if local correlations dominate therenormalization in the band, it is highly plausible that thesame considerations apply elsewhere in the Brillouin zoneas well.We are able to investigate this issue by assuming that thewhole-bandwidth renormalization applies throughout theCarrier number (per Ru)0.84.44.3 a)4.24.14.03.9 Total electron~ ~1.41.21.00.80.60.4y = 0.2710 K(electron)(electron)ARPESdHvATight-Bindinge)0.8(hole)0.2(hole)00 0.1 0.2Additional electron y0.3y = 0y = 0 (100 K)y = 0.18y = 0.20y = 0.27FIG. 2 (color online). Scans along the -M- line in the Brillouin zone for samples with y 0, y 0:18 and y 0:27 at 10 K andE


asic structure of a lab reporttotal length < 2500 words (not incl. abstract)abstract (< 200 words)introduction (< 400 words)theoretical background (< 500 words)apparatus & description of experimentwith diagram (< 500 words)analysis & discussion of results(AT LEAST 1000 words)conclusion (< 300 words)• abstract should includemain results, relevant #s• figures (3-5 in paper) arethe most important part!Figures + captions shouldbe self explanatory• analysis and discussionis 2 nd most important.Can subdivide if yourexperiment has multipleparts


what NOT to do (a typical first lab report)abstractintroductiontheoretical backgroundapparatus & description of experimentanalysis & discussion of resultserror analysis sectionconclusion (< 300 words)• common when one doesnot work on an outline /rough draft• theoretical part is usuallyeasy to write (regurgitation)but also least important• having an inadequateanalysis section is a surefireway towards a poor grade• do not write a separateerror analysis section


asic do’s and don’tsDO• get started early and work on a rough draft (I like to use apad of Post-It notes) to figure out what you want to say• spend a serious amount of time working on your figuresand making them as clear & attractive as possible• spend the majority of your effort on the data analysis andfocus on discussing the physics of what is happening• include error bars and uncertainties in your analysis• try to use a “serious” typesetting program such as LaTeX( various front-ends exist : TeXShop (Mac), LyX (Mac/Win) )• feel free to have your instructor comment on your drafts


asic do’s and don’tsDON’T• let your lab report get too long! (main body < 2500 words)• spend too much time regurgitating theoretical derivations.Keep the theory section short and sweet ( < 500 words)• write a separate “error analysis” section. Discussion ofuncertainty should be embedded in data & analysis sections• worry too much about getting a preconceived / acceptedvalue, although you should still compare to previous work.• give more significant digits than your uncertainty estimatesallow. NO : 1.3334 ± 0.3321 eV YES : 1.3 ± 0.3 eV• get started the night before! Preparation shows!


eferences and attribution• You must include references your sources andinclude attribution• Wikipedia is not sufficient for most material: reference(and read) original sources• Copy-paste is never acceptable


an example abstract (Stern-Gerlach)Goodlength is appropriatesuccinctProblemsno numerical results mentioned(should stand alone)The abstract may be the only thing that most people willread. It should be self-contained and hook people in


an example introduction (muon lifetime)Goodsuccinct (this is not entire intro)interesting historical context withrelevant informationattributions (references)


an example theory section (Stern-Gerlach)Problemslong and overly pedantic, lots of unnecessary detail (almost all of it!)ripped directly off Wikipedia (no attribution!)


an example apparatus figure (optical pumping)Goodclear schematic viewProblemscaption could be moreinformativelabels could be more detailed“filter 1”


an example data plot (Stern-Gerlach)Goodnice use of colorProblemstoo smallno error barsno units, missing y-axis labelcaption is incomplete


an example data plot (from published PRL)PRL 99, 187001 (2007)PHYSICAL REVIEW LETTERSweek ending2 NOVEMBER 2007Energy (meV)0-50-100a)0.8y = 010 K1.01.2b)0.8y = 0100 K1.01.2c)0.8y = 0.1810 K1.01.2k x(π/a)d)0.8y = 0.2710 K1.01.2e)0.8y = 0y = 0 (100 K)y = 0.18y = 0.20y = 0.271.0y = 0.27y = 0.20y = 0.18y = 01.220100-10-20FIG. 2 (color online). Scans along the -M- line in the Brillouin zone for samples with y 0, y 0:18 and y 0:27 at 10 K andE


an example conclusion (diffraction)ProblemsVague! What improvements?Conclusion should state clearly “using technique A we measuredx ± y which demonstrated within our uncertainty that....”


some common pitfalls• including too much in the report (keep it under 2500 words!!!)• quoting the uncertainty in your measurement as the difference between yourresult and the accepted value for the parameter that you had measured• copying a long theoretical treatment from a book ...– only include the pertinent background comments and formulae– and not properly referencing it• forgetting to include error bars on the graphs or using inappropriatesignificant figures• omitting an appropriate diagram of the apparatus• too many detailed diagrams of the circuits, layout, pictures, etc... of theapparatus• including material that has little to do with the experiment that you havecarried out


some final advice• figures are the most important part of a paper.– Make them clear, attractive, properly labeled. Use descriptive figurecaptions, properly numbered. Graphs should typically include error barson the data points• writing a good report takes time– starting the analysis the night before is not a successful strategy. Makesure you have all your key data or parameters well in advance– starting the analysis while you are taking the data is a very useful strategy• begin writing the report as you are carrying out the final analysis– writing an outline and/or rough draft is incredibly helpful– let the report sit for a day, reread it and then correct its shortcomings• The primary goal of the report is not to compare to previous value.– It is to report on the experiment you did and estimate the accuracy of theresults WITHOUT CONSIDERING THE EXPECTED OUTCOME.

More magazines by this user
Similar magazines