1. What are we talking about? Naïve picture:Δx ↕ ~
external drive Ω
fixed suspension
mass M
M ~ 10–18 kg (NEMS) ~ 10–21 kg (C nanotube)
Ω ~ 2π ∙ 108 Hz
Teq ≡  Ω / kB ~ 5 mK , x0 ~ 10–12 m
rms groundstate displacement
Actually:s
↕ Δ x d
In practice, Δx « d.

2. DOES VANISHING OF EVIDENCE LEGITIMATE REINTERPRETATION OF MEANING?
WHAT WILL WE LEARN?
If QM predictions obeyed:
check QM to “more macroscopic” scales? 
learn mechanisms of decoherence? 
“understand QM-classical transition”? NO!
______________
No doubt we will, eventually, see effects of decoherence in experiments. BUT:
Decoherence does NOT solve the “measurement” (realization) problem!
At micro-level:
QM description α |1  + β |2  cannot be interpreted as “each microsystem is either |1  or |2 ”
Evidence: interference
At macro-level:
QM description is still α |1  + β |2  (formalism is unchanged) but interference destroyed by decoherence.
DOES VANISHING OF EVIDENCE LEGITIMATE REINTERPRETATION OF MEANING?

3. WHAT WILL WE LEARN? (cont.)
(B) IF QM PREDICTIONS FAIL: (or we contemplate the possibility that they may fail)
Test QM against
— classical mechanics
— specific theories of realization (e.g. GRWP, Penrose)
— general class of macrorealistic (“MR”) theories.
Some tests already done with other systems (fullerenes, quantum-optical systems, SQUIDS)
in SQUID experiments, check QM predictions for superpositions of states in which large number N of microscopic entities (e–) behave differently (N~104–1010).
What are advantages/disadvantages of MEMS vis-à-vis SQUIDS?

4. SIMPLE HARMONIC OSCILLATORS!
MEMS vs SQUIDS
SQUID:
Ψ~ a|↺+ b|↻
I ~ 1-5 μA, N ~ 104–1010
(a) fidelity of readout quite poor (but rapidly improving)
(b) states |↺, |↻ do not differ in distribution of center of mass (trigger for GRWP, Penrose . . .)
 cannot discriminate between these particular MR theories and QM (but can test more general class of MR theories)
In these respects MEMS much superior. But: major disadvantage of MEMS: to a first approximation,
SIMPLE HARMONIC OSCILLATORS!
Under driving force F(t), behavior of mean oscillator coordinate <x>(t) identical in QM and classical mechanics!

5. HOW TO GET AROUND QUANTUM-CLASSICAL CORRESPONDENCE FOR SHO?
Couple to quantum-limit (2-state) system so as to generate strongly nonclassical states
Ψ(x) = αΨ1(x)+ βΨ2(x)
(2-state system: photon, trapped ion. . .)
Introduce anharmonicity + external drive, study bistable behavior (R. Lifshitz)
General problem in testing QM vs. “physical collapse” theories (e.g. GRWP, Penrose):
DECOHERENCE MIMICS EFFECTS OF COLLAPSE!
So, want simultaneously:
“large” GRWP (etc.) collapse rate (coll)
“small” decoherence rate (dec)
∆ : under experimentally realistic conditions (Δx « d)
coll ~ Δx (GRWP), ~ (Δx)2 (Penrose)
dec ~ (Δx)2  don’t gain by increasing Δx .
–Ψ1
–Ψ2

6. HOW CONFIDENT ARE WE ABOUT (STANDARD QM’l) DECOHERENCE RATE?
Theory:
(a) model environment by oscillator bath (may be questionable)
(b) Eliminate environment by standard Feynman-Vernon type calculation (seems foolproof)
Result (for SHO):
ARE WE SURE THIS IS RIGHT?
Tested (to an extent) in cavity QED: never tested (?) on MEMS.
Fairly urgent priority!
provided kBT»Ω
energy relaxation rate (Ω/Q)
zero-point rms displacement