This is an annotated list of important publications. For an exhaustive list, try an ADS author query.

### BICEP/Keck results papers

The primary goal of the BICEP/Keck series of experiments is to search for evidence of inflation / primordial gravitational waves (PGW) using B-mode polarization of the CMB.
The PGW amplitude is parametrized by tensor-to-scalar ratio, *r*.
Our sensitivity to *r* has improved over time by making deeper maps, increasing our frequency coverage to improve separation of CMB from foregrounds, and continuing control over instrumental systematics.

#### BICEP/Keck XIII: Improved Contraints on Primordial Gravitational Waves using Planck, WMAP, and BICEP/Keck Observations through the 2018 Observing Season, Phys. Rev. Lett. 127, 151301, 2021, arXiv:2110.00483

This paper incorporates new 95 GHz data from BICEP3 and 220 GHz data from Keck Array (2016–2018).
This results a big step forward in sensitivity to *r*, with the one sigma error decreasing from 0.02 to 0.009.
Our simple parametrized mode of CMB + foregrounds continues to be a good description of the data.
With this improvement in sensitivity, the combination of BICEP/Keck constraints on *r* with Planck constraints on *n _{s}* rules out the full classes of monomial inflation models and natural inflation (see Fig 5).

#### BICEP/Keck X: Constraints on Primordial Gravitational Waves Using Planck, WMAP, and New BICEP2/Keck Observations through the 2015 Season, Phys. Rev. Lett. 121, 221301, 2018, arXiv:1810.05216

This paper includes our first 220 GHz data, taken with Keck Array in 2015.
The new high frequency data improves our ability to identify dust and decreases the one sigma error on *r* from 0.024 to 0.020.
In Appendix F we explore an extension to our foreground model, adding a dust decorrelation parameter to handle spatial variations in the dust spectral energy distribution.
There is no evidence for dust decorrelation at the current level of sensitivity, but we expect that it must be present at some level due to the complexity of the interstellar medium.

#### BICEP/Keck VI: Improved Constraints On Cosmology and Foregrounds When Adding 95 GHz Data From Keck Array, Phys. Rev. Lett. 116, 031302, 2016, arXiv:1510.09217

For the 2014 season two of the five Keck Array telescopes were converted to observe at 95 GHz, the first of several frequency upgrades for Keck Array.
This frequency band has lower levels of polarized dust contamination than the 150 GHz band that was used for all previous BICEP2 and Keck Array observations.
After one year of operation, the depth of these new maps in the BICEP field greatly exceeded that of the Planck 100 GHz map, leading to an improved upper limit *r* < 0.09 at 95% confidence and significantly better discrimination between dust and CMB signals.

#### BICEP/Keck and Planck Collaborations — A Joint Analysis of BICEP2/Keck Array and Planck Data, Phys. Rev. Lett. 114, 101301, 2015, arXiv:1502.00612

In this paper we joined forces with the Planck Collaboration, who had high frequency (353 GHz) data that revealed polarized dust in the BICEP field.
This paper marked the first appearance of our parametrized multi-component likelihood, which uses all of the auto and cross spectra from a set of maps (in this case, the BICEP/Keck 150 GHz map plus Planck maps at seven frequencies) to constrain a model of cosmology plus foregrounds.
With the help of the multi-frequency Planck data, this analysis showed that the BICEP2 signal was entirely consistent with dust from the Milky Way and we set an upper limit *r* < 0.12 at 95% confidence.

#### BICEP/Keck I: Detection of B-mode Polarization at Degree Angular Scales by BICEP2, Phys. Rev. Lett. 112, 241101, 2014, arXiv:1403.3985

The first BICEP2 results paper and the first detection of B modes at degree angular scales. At the time this paper was first published our internal analysis (using BICEP1 data) gave a weak hint that the detected signal had CMB origin and the models of galactic dust that we considered were all much smaller. As it turned out the models were overly optimistic about the degree of dust polarization at high galactic latitudes, which we learned once Planck published their first results.

### BICEP/Keck systematics papers

#### BICEP/Keck XI: Beam Characterization and Temperature-to-Polarization Leakage in the BK15 Dataset, Astrophys. J. 844, 114, 2019, arXiv:1904.01640

This is a companion paper to the BK15 B-mode results paper (BK-X).
It includes a summary of beam map data for the Keck Array 2012–2015 seasons and simulations of the expected temperature-to-polarization leakage from higher order modes after deprojection.
In Section 6 of the paper, Cincinnati postdoc Caterina Umiltà led an exploration of how to best propagate the simulated error to our constraint on *r*.
These methods included evaluating the bias incurred for several estimates of the leakage as well as including a leakage power template with amplitude parameter in the likelihood analysis.

#### BICEP/Keck IV: Optical Characterization and Performance of the BICEP2 and Keck Array Experiments, Astrophys. J. 806, 206, 2015, arXiv:1502.00596

In this paper, we present high signal-to-noise beam maps obtained *in situ* for BICEP2 and the Keck Array 2012 and 2013 configurations.
To obtain the beam map data, we built a chopped thermal source and operated it on a mast at the South Pole; large flat mirrors were used to redirect the BICEP2 and Keck Array lines of sight over their respective groundshields to see the source.
From the beam maps, we can predict the magnitude of temperature-to-polarization leakage systematics, though the systematics deprojection procedure works without this input.
This paper also includes measurements of the BICEP2 polarized beam, including absolute polarization angles.

#### BICEP/Keck III: Instrumental Systematics, Astrophys. J. 814, 110, 2015, arXiv:1502.00608

This paper contains a thorough evaluation of instrumental systematic errors affecting the BICEP2 B-mode result.
An important innovation described here is the method of *systematics deprojection*.
When paired detectors have mismatched responsivity or beam shapes, these differences cause temperature-to-polarization leakage.
But we have good knowledge of the CMB temperature map from Planck, so we can calculate the specific patterns in the map caused by the leading order systematics (monopole, dipole, quadrupole) and then regress these modes out of our data.
This technique was key to producing a reliable BICEP2 result, as well as all future results from the BICEP/Keck program.

### CMB-S4 forecasting and white papers

CMB-S4 is a next-generation array of telescopes to study the Cosmic Microwave Background with unprecedented sensitivity. This project is in the design and planning stages and will be built at two sites: the South Pole and the Chilean Atacama desert.

#### CMB-S4: Forecasting Constraints on Primordial Gravitational Waves, Astrophys. J. 926, 54, 2022, arXiv:2008.12619

A key science goal of CMB-S4 is to either detect primordial gravitational waves if tensor-to-scalar ratio *r* is greater than 0.003, or else to set a 95% confidence upper limit of *r* < 0.001.
In this paper, we describe a framework for forecasting CMB-S4 sensitivity to *r* that is grounded in the achieved performance of past and current experiments like BICEP2, Keck Array, and BICEP3.
This framework allows us to iteratively develop designs for the CMB-S4 small aperture telescopes, estimate their sensitivity, and then check the results through analysis of simulations.
The CMB-S4 Low-ℓ BB Analysis Working Group continues to use this methodology for forecasting.