Quantum Communication Primitives

Implementations of the three foundational quantum-communication protocols in Qiskit, each with a self-checking verification so you can confirm the physics actually works — not just that the code runs.

Protocol	What it does	Resource trade
Quantum teleportation	Moves an unknown qubit state from Alice to Bob	1 Bell pair + 2 classical bits
Superdense coding	Sends 2 classical bits by transmitting 1 qubit	1 Bell pair + 1 qubit
BB84 QKD	Distributes a secret key; detects eavesdroppers	quantum channel + public sifting

Teleportation and superdense coding are duals of each other — one trades a shared Bell pair to send quantum information with classical bits, the other to send classical information with a qubit. BB84 turns the same measurement-disturbance physics into a security guarantee.

Results

Teleportation — a random single-qubit state is teleported and verified by un-preparing it on Bob's side; a perfect transfer returns his qubit to |0>:

Bob's qubit returned to |0> in 4096/4096 shots (success = 1.0000)
PASS: arbitrary state teleported with fidelity > 0.99

Superdense coding — all four 2-bit messages delivered through a single qubit:

message 00 -> decoded 00   message 01 -> decoded 01
message 10 -> decoded 10   message 11 -> decoded 11
PASS: 2 classical bits delivered per transmitted qubit, all messages.

BB84 — with no eavesdropper the keys match exactly (QBER ≈ 0); an intercept-resend eavesdropper is exposed by the error rate jumping to ~25%:

No Eve:    sifted key 303 bits, QBER 0.000, keys identical
With Eve:  sifted key 303 bits, QBER 0.254  -> attack detected

Running on real quantum hardware

The protocols above run on a perfect, noiseless simulator. To see what real qubits actually do, src/run_on_hardware.py runs a Bell pair on an IBM superconducting quantum processor and compares it to the ideal simulator.

A Bell pair should produce only 00 and 11 (the two qubits always agree). On the real 156-qubit ibm_fez (Heron r2) processor:

Outcome	Ideal simulator	Real hardware (ibm_fez)
00	1021	972
11	1027	992
01	0	32
10	0	52

The entanglement survives 95.9% of the time; the 01/10 outcomes — which are physically forbidden for an ideal Bell pair — appear at a 4.1% error rate from real device noise (readout error, two-qubit-gate error, and T₁/T₂ decoherence). The 10 > 01 asymmetry is consistent with T₁ relaxation of the excited state plus qubit-specific readout error.

# requires a free IBM Quantum account (credentials saved once)
python src/run_on_hardware.py

Run it

pip install -r requirements.txt
python src/teleportation.py
python src/superdense_coding.py
python src/bb84_qkd.py

How each protocol works

Teleportation (src/teleportation.py)

1. Prepare the unknown state |psi> on Alice's payload qubit.
2. Share a Bell pair between Alice and Bob.
3. Alice does a Bell-basis measurement (CX, H, then measure) on her two qubits.
4. Bob applies X and/or Z corrections conditioned on Alice's two classical bits.
5. Verification: applying U† on Bob's qubit returns it to |0> iff teleport succeeded.

Superdense coding (src/superdense_coding.py)

Alice's local gate (I, X, Z, or ZX) on her half of a Bell pair rotates it to one of the four orthogonal Bell states; Bob distinguishes all four with a Bell measurement, recovering both bits.

BB84 (src/bb84_qkd.py)

Each bit is encoded in a random basis (Z or X). Eve cannot copy an unknown qubit (no-cloning), so her intercept-resend guess in the wrong basis injects ~25% errors on the sifted key — making eavesdropping statistically detectable.

Notes

Built while studying quantum computing (CDAC / IIT Roorkee / MeitY course, 2025). Tested on Qiskit 2.4, Python 3.12.