Interesting (and doable) Thesis Topics

---

The `doability' needs to be examined in a feasibility study before anybody gets comitted to anything.  Also, the `doability' is a function of time.  What is clearly not doable now, might be in a year, and certainly could be in two.  It all depends on your time-scale.

There certainly are really cool topics I forgot about or haven't even heard of.  Let me know and I'll add them to this list.  I'm a personal fan of B physics (at least this early in Run 2a), but that shouldn't constrain you.  What you should be looking for is the right mixture of coolness and feasibility.
 

Flavor tagging:

There was already a lot of discussion (both oral and via email) about this, but let me nonetheless reiterate that in my view the reconstruction of the slow B measons and extending the lepton coverage to the high-eta region are not only efforts that will be exciting and fun to do, but also very important for both the CDF and our group, and perhaps even for your own career, since they can be build upon and thus recycled for other analyses.  (Just think of tri-lepton searches for SUSY, or even  , or SUSY channels involving charm + MET.  Get the picture?)  The flavor tagging in fact is just an excuse (or a stimulant) to get these tools working.  The flavor tagging itself is the most valuable contribution to the  mixing and CP violation projects -- and these taggers, once they work, should contribute heftily -- however any thesis that revolves around these tools (including SUSY) is fine.  As I said a dozen times before, what's valuable are the tools -- the ability to reconstruct soft leptons and b's -- the actual thesis project is less critical.  (Although going for a lesser thesis would be beyond me.)

The flavor taggers in this picture are:

• vertex charge flavor tagging
• soft lepton tagging at high eta (ele and mu)
• tagging
• NEW: Same Side Tagging

Vertex Charge Tagging (or VQT for short) is an extension of the `standard' Jet Charge Tagging (JQT) as developed for Run I.  The problems with the Run I approach are that:

• it was developed for the higher momentum B's (say ~ 20 GeV/c on average)
• it was hard to find the opposite side jet and be sure that that's a b jet (i.e., lots of gluon jets as well -- the NLO contribution to the  production is larger than the LO, so gluons predominantly figure in the final state; luckily they are usually soft...)
• in Run II, the track-jet clustering may not work due to low momentum of the opposite side b-jet, which would make the b daughters less colimated (and in fact with very slow b's, the daughters would go in all directions!)

1. 1: just find a displaced vertex on the opposite side, and use that information to guide the track selection and momentum weighting
2. 2: after the opp.side displaced vertex has been found, try to isolate the b-daughters from the primary tracks and weigh them differently; the primary tracks are treated the same as in the `Same-Side Tagging' analysis (which is, as the name says, done on the side of the B meson to be tagged :)
3. 3: finally, try to separate the secondary vertex into the B and D components, and use that information as an extra handle on the decay flavor

Moreover, any improvements of the silicon tracking -- like making the L00 work or measuring the resolution of the 90-degree strips -- naturally feed into this analysis.  So there's a lot of synergy with other efforts in our group.

The results of these studies can be applied elsewhere in a number of ways, from searches for slow heavy flavor for SUSY, to charm counting to `b-hadron veto'...   In B physics, one can use the capability to add other decay products to an incomplete secondary vertex in order to do semi-exclusive decay reconstruction.  For instance, one can do a  mixing analysis with  X where `X' is a collection of likely  daughter tracks that has been associated with the found  vertex.  (I've done this in Run I but gave up since the algorithms to associate tracks to vertices were non-existent and the payoff to develop them wasn't there.  But now we have TTT and already have thousands of  's).
 

Soft Lepton Tagging (SLT) at high- is obviously useful on a multitude of levels, including the SUSY trilepton analyses that should also involve the plug.  The only show-stopper is the (lack of) standalone tracking.  We need to decide whether to wait until it's ready, or jump in and try to fix it ourselves.  Other than that, it's a non-trivial but relatively straightforward process: electrons are reconstructed by a combination of PES/PEM/PPR and the tracking info (momentum) and muons by IMU/BMU combination.  Yi has already poured a lot of time into the former, while Satyajit looked into the latter and it looks promissing.

Obviously, any analysis that needs slow leptons in the plug can use the tools and techniques developped for SLT.  In the B world there are several, and I list a few below.  One not mentioned is the analysis of  correlations at high- which is important for both B-physics and the Higgs searches (since there the  X  production is the largest background).  (Note that the slow-b-finding can also be used for the same purpose.)
 

Flavor tagging using opposite side  is really similar to VQT, except that it's far simpler.  In fact it's so simple it's probably more appropriate as an undergraduate thesis, and I list it here simply because it might be a fairly quick exit option for somebody like Rob.   The idea is to use a fully reconstructed decay on the opposite (i.e. flavor tagging) side to tell the flavor of the opposite b-hadron.   One can also use  decay since there it's pretty obvious which track is the proton, and its charge carries the information of the flavor of the other b-hadron.

Again, an alternate thesis option here is to do a SUSY search with D's and MET, but we should ask around whether there are Exotics students working on it.
 

NEW: Same Side Tagging (or "SST"), study of  (orbitally excited B mesons)

I did the SST as my own Ph.D. so I'm not terribly excited about supervising somebody else.  However, this is an important study that has to be done, and I definitely have the required expertise.  I'm adding this topic here since, thanks to the Lepton + SVT trigger, as of Aug 2002 we appear to have about 70% of the Run I semileptonic statistics, and that makes serious studies of the origin of of the SST effect (that is, the correlation of the flavor of the b-meson and the charges of particles in a cone around it) at least feasible, with a promise of enough statistics by the summer of 2003 to make it conclusive.

(Bonus: in Run I, we used an algorithm which hasn't been completely optimized.   I know what needs to be done to make it better...)

(Another bonus is the synergy with VQT, since the tracks assigned to the primary vertex by VQT (with certain probability) are treated as fragmentation tracks and thus exhibit at least some SST correlation to b-hadron flavor...)
 
 

Systematic study of Penguins:

The  decays, where the photon is either virtual (thus giving rise to two oppositely charged leptons) or real (and is reconstructed either as a photon by the calorimeter or as a conversion pair) are extremely interesting for two reasons:

• independent measurement of  -- the  mixing is then used to compare them and look for contributions beyond the Standard Model
• high rates of these rare decays (i.e., beyond the SM predictions) can by themselves indicate non-SM effects

Here are a few interesting Penguin decays:
 

The gluonic penguin    (experimental problem: no good trigger).  This is a pure penguin decay.  Measuring CP asymmetry gives sin2beta in the Standard Model, however if there are New Physics contributions, one may expect the three and the penguin to be affected differently.  Thus a statistically significant discrepancy is a sign of the New Physics.

Experimentally, the biggest problem for this decay is the lack of good trigger at CDF.  The only reasonable try is TTT however the two kaons from the phi are often not resolved by the SVT and thus the trigger efficiency for this decay is not as high as for most of the usual B decays.  However, the decays    and    are not troubled by this since there are 3 and 4 tracks respectively emerging from the B vertex.  All three have branching ratios at the level of  .  But BELLE already sees about 10 evts, so as long as we beat that, this is interesting.

There are two ways to approach  :

• rely on TTT and get what we can; from earlier studies by Mat M., we know that SVT passes the two kaons from  in about 10-20% of the cases.
• we can use lepton + track trigger sample, and require the lepton to be from the other b-hadron on the opposite side.  But here we pay the price of the coverage on the opposite side and semileptonic branching ratios of the other b.

EM penguins with a lepton pair in the final state

Here we are after the decay  (where Xs is a strange-flavored meson).  Two parameters are of interest:

• the rate (that is, the B branching ratio)
• the forward-backward asymmetry,  , is zero in SM but very sensitive to the New Physics (especially SUSY).

• one or both muons at high-   -- IMU/BMU extend the coverage and that may play an important role in the forward-backward part of this analysis
• use central electrons (CES, etc.)
• use plug electrons (PES, etc.)

EM penguins with photons and conversions

Photons: needs a trigger which we had in Run I, and which, in some form, made it into the trigger table.  However the resolution of the calorimeter is not very good which makes the B mass peak rather broad, which makes this approach not very effective.

That is rectivied when the photons are reconstructed as conversions.  With conversions, one can not only go to lower energies, but also the tracking information provides for a really good photon energy resolution, and thus with a dramatically improved B mass resolution.
Moreover, there are these huge conversion samples, and we might even have a `conversion trigger' (which was at least talked about some time ago).  Somebody needs to make a back-of-the-envelope calculations of the possible yields (or check those provided by Barry W. and Masa when they proposed the trigger years ago), and then we talk more.

Note that the Universal Finder supports combining the conversion candidates (from VertexFit) with other particles to form a decay tree.  Thus technically the code for this analysis has been written.  But there's still a lot of physics to do.
 
 

NEW!!!  SUSY thesis:

Anomaly-mediated SUSY breaking is a new flavor of SUSY that fixes some troubling spots of vanilla SUSY (e.g. the hierarchy problem) and is thus one of the theorists' recent favorites (e.g. Raman Sundrum's).  In this model the neutralino is LSP (say around 150 GeV) and chargino is just above it.   Chargino decays into the neutralino plus a few slow pions (so the signature is lots of MET + a few low-Pt tracks) and due to low mass difference the lifetime can be
very long -- sometimes milimeters, centimeters and even meters.   While the displaced vertex in the middle of COT is a bit hard to reconstruct (requires special tracking), a chargino decaying inside the beam pipe is a fair game.

Apparently nobody looked at this in Run 1, presumably because everybody assumed that only displaced vertices will arise from heavy flavor decays, and it will be in jets.  There are no jets here: just lots of MET and a few low momentum particles going in all directions (or in one direction by not radially from the beamline).  

This type of geometry is ideally suited for our TopologicalVertexFinding algorithm we designed and optimized for slow B mesons.  Except that this analysis is done on the MET sample, not lepton + SVT.

NEW!!!  Preparatory theses:

These theses aren't so hot and useful by themselves, but they deepen our understanding of the important measurements to come.  An excellent example of such analysis were the measurements of two-body charmless B decays performed by CLEO between 1997 and 1999,
in which the decays like  and  were observed for the first time and their branching ratios measured.  Neither CDF nor the B-factories had any idea how well their CP asymmetry in    are going to work out since nobody knew the branching ratio, apart from the theoretical expectations.   So here a few examples in that vein:
 

Observation of  and the measurement of BR( )/BR( )

The final state in the decay  is not a CP eigenstate, but the total amplitude is a sum of two tree-level amplitudes.  The Cabibbo suppressed decay features Vub and thus that amplitude picks a phase shift of  , resulting in an interference and thus CP violation.  However, here the time-dependent CP asymmetry measures sin .  The problem is that nobody really knows how much smaller the rate for  is (as compared to the flagship  )  and what the S/B in that sample is.  Furthermore, the background from the latter is severe (we guess about 20 times the signal), and therefore the analysis crucially depends on dE/dx since the success hinges on being able to distinguish kaons from pions.

My former student Stephen Bailey studied this for the Harvard equivallent of GBO, and the study ended up in the Yellow book.  The bottom line is that we might expect about 600-700 events for 2/fb, and thus for 100-200/pb there could be as much as 30-70 signal events.  Depending on the background, this could be > 4 sigma signal and thus at least a (very!) useful measurement of BR( ).
 
 

NEW!!!  "Lesser" theses:

In this section I'll offer severeal other topics that are interesting, fun, and doable on a short time-scale but perhaps not as relevant for the
grand picture as the flavor tagging analyses.  Here they are:
 

Measurement of  and  mass spectra

The main interest for the orbitally excited (L=1) B and D mesons is motivated by the Same Side Tagging.  's are also a significant physics background in various lepton + D analyses which require separation of charged and neutral B's.  The HQET predicts the various features of the mass spectra with various level of trustworthiness (e.g. mass splitting between various states are better known than the actual location of the center of the spectrum).  Also B and D spectra are similar (the quarks are heavy) so from their comparison we can learn something about the quality of HQET predictions.

Since there are two narrow and two wide states, the analysis can be tricky, but I think it's doable.  Experimentally, the charm part should be a breeze, since there are tens of thousands of fully reconstructed charm states in the TTT, and a large fraction (say 30% although nobody really knows :) come from the higher states.  B's are harder, but there may be enough fully reconstructed B's by the next summer to make this into a worthwhile project.
 

Measurement of various  branching ratios

The only exclusive  decay reconstructed so far is  .  However, there should be several exclusive decay modes that can be reconstructed in the TTT sample, starting from  , then  and so forth, down the PDG list...  These are interesting for a couple of reasons:

• we would like to know what these branching ratios are
• we can use these decays as a springboard for reconstructing  , e.g. in channels like 
• we can learn a lot about HQET and the factorization by comparing  to 

Good factorization tests:

It is necessary to use fact. in several important constraints on the CKM matrix, so developing analyses that can probe the factorization hypothesis (and provide direct measure of how much the Nature deviates from it) is very useful.

The measurement of   (here neutral kaon becomes  or  ).  This decay mode is a pure penguin.  In this analysis we have to rely on BaBar/BELLE to measure     (which is a pure tree, but beyond our experimental capability).  One need both to say anything about the factorization based on the comparison of these two topologically similar decays...
 

Comparison of  to 

See above.
 
 

Direct CP violation:

More coming here later.