---
title: 'Syn2Co: Dual Applications in Vision & Surface Science'
url: https://www.emergentmind.com/topics/syn2co
type: topic
---

# Syn2Co: Dual Applications in Vision & Surface Science

Syn2Co has been used in two distinct ways in recent arXiv literature. In self-supervised computer vision, it denotes **Synthetic-to-Contrast**, a framework that combines synthetic data augmentation in image space with synthetic hard negatives in representation space for Vision Transformer pre-training on ImageNet-100 [2509.02029]. In surface chemistry and 2D conjugated polymers, it denotes a protocol for selective on-surface 2D covalent polymerization of fluorographdiyne nanosheets on Au(111) through the combination of cobalt catalysis and coronene templating [2606.00495]. The shared label does not imply a shared methodology; the two usages address unrelated technical problems and operate at different levels of abstraction.

## 1. Nomenclature and domain-specific meanings

The term is best understood as an acronym with domain-dependent semantics rather than as a single canonical method.

| Usage of “Syn2Co” | Domain | Core meaning |
|---|---|---|
| Synthetic-to-Contrast | Self-supervised vision | Synthetic data plus synthetic hard negatives in a momentum-encoder InfoNCE pipeline |
| Syn2Co protocol | Surface synthesis | Cobalt catalysis plus coronene templating for fluorographdiyne growth on Au(111) |

In the vision usage, Syn2Co is explicitly expanded as **Synthetic-to-Contrast** and is designed for contrastive self-supervised learning with DeiT-S and Swin-T backbones [2509.02029]. In the surface-science usage, Syn2Co refers to an on-surface polymerization protocol that synthesizes single-layered fluorographdiyne nanosheets up to \(60\times 60\ \mathrm{nm}^2\) on Au(111) [2606.00495].

A common source of confusion is that acronym-based search can also retrieve the nearby term **SyCo**, short for **Synthetic Coordinate Embedding**, a molecular graph generation framework in latent Euclidean space rather than a Syn2Co method [2406.10513]. This suggests that citation by arXiv identifier is especially important when the acronym alone is ambiguous.

## 2. Syn2Co in self-supervised vision

In the computer-vision literature, Syn2Co targets the standard contrastive objective of learning an encoder \(f:X\to\mathbb{R}^d\) such that positive pairs are nearby in feature space while negatives are far apart. Its defining move is to combine two “faking” strategies: **synthetic data augmentation in image space** and **synthetic hard negatives in feature space**, both inserted into a momentum-encoder, queue-based InfoNCE pipeline for Vision Transformers [2509.02029].

The synthetic-data component uses a class-conditional diffusion model, **Relay Diffusion [22]**, trained to clone **ImageNet-100’s class distribution**. The resulting synthetic set is \(X_s \approx 130{,}000\) images, stated as **one per original ImageNet-100 sample**. Pre-training then mixes real and synthetic images by sampling a batch in which a fraction \(\alpha\in[0,1]\) comes from real data \(X\) and the remainder \(1-\alpha\) from \(X_s\). The boundary cases are explicit: \(\alpha=1.0\) is pure real pre-training and \(\alpha=0.0\) is pure synthetic pre-training.

The synthetic-negative component starts from a standard momentum-encoder pair \((f_q,f_k)\), temperature \(\tau\), and a queue \(Q\) of size \(K\). For each query \(q=f_q(x_q)\), cosine similarities \(\mathrm{sim}(q,n)=q^\top n/\|q\|\|n\|\) are computed for all \(n\in Q\), after which the top-\(N\) hardest real negatives are selected. Synthetic negatives are then produced by applying a synthesis function \(F(q,n)\) to these hard negatives. The paper lists **six synthesis modes from SynCo [12]**, including **interpolation**, **extrapolation**, **mixing**, **jittering**, **perturbation**, and **adversarial variants**. Each synthetic negative is normalized as \(s=F(q,n)/\|F(q,n)\|\), and the final negative pool becomes the union of real and synthetic negatives.

The resulting loss augments the usual InfoNCE denominator with synthetic negatives:
\[
\mathcal{L}_{\mathrm{Syn2Co}}(x)
=
-\log
\frac{\exp\bigl(q\cdot k^+/\tau\bigr)}
{\exp\bigl(q\cdot k^+/\tau\bigr)
+\sum_{n\in N_{\mathrm{real}}}\exp\bigl(q\cdot n/\tau\bigr)
+\sum_{\tilde s\in N_{\mathrm{syn}}}\exp\bigl(q\cdot \tilde s/\tau\bigr)}.
\]

The architecture scope is narrow and explicit: **DeiT-Small (21 M parameters)** and **Swin-Tiny (29 M parameters)**. The evaluation protocol is equally explicit: **linear probing**, with the encoder frozen and a single linear classifier trained for **100 epochs**, reporting **top-1/top-5 accuracy on the ImageNet-100 validation set**.

## 3. Training dynamics, hyperparameters, and empirical behavior in vision

The training loop samples \(B_{\text{real}}=\lfloor \alpha B\rfloor\) from \(X\) and \(B_{\text{syn}}=B-B_{\text{real}}\) from \(X_s\), applies two augmentations per image, computes query and key embeddings, mines the top-\(N\) hard negatives from the queue, synthesizes \(L\leq N\) feature-space negatives, evaluates \(\mathcal{L}_{\mathrm{Syn2Co}}\), updates the online encoder, performs the momentum update \(\theta_k\leftarrow m\theta_k+(1-m)\theta_q\), and refreshes the queue [2509.02029]. The key hyperparameters listed are \(\alpha\), \(K\), \(N\), \(\rho=L/K\), \(\tau\), \(m\), batch size \(B\), and epochs \(E\). Representative values in the paper include \(K=65536\), \(N=256\), \(\rho=0.1\text{–}0.25\), \(\tau=0.1\text{–}0.2\), \(m=0.999\), batch sizes \(128\text{–}2048\), and \(E=100\text{–}200\).

The reported linear-probe results on ImageNet-100 show that the gains are architecture-dependent.

| Method | DeiT-S Top-1 | Swin-T Top-1 |
|---|---:|---:|
| DINO | 79.41 | 81.78 |
| MoBY | 79.36 | 83.90 |
| Syn1Co (data only, 152 400) | 81.86 | 83.68 |
| Syn1Co-Neg | 78.96 | 84.04 |
| Syn2Co (full, 152 400) | 82.12 | 83.70 |

The main quantitative highlight is that **DeiT-S gains \(+2.76\%\) Top-1 over MoBY when using both synthetic data and synthetic negatives at 200 epochs**. For **Swin-T**, the paper reports **modest gains (\(\sim +0.14\%\)) from synthetic negatives alone**, while **synthetic data adds limited benefit**. The study also reports that a **purely synthetic pre-training run (\(\alpha=0\)) is within \(\sim 5\text{–}10\%\) of the fully real regime** in linear-probe accuracy, using this as an empirical proxy for diversity.

The paper’s own analysis is cautious. It states that low-resource regimes benefit more from mixing in diffusion-generated images, that extended training extracts more signal from synthetic clones, and that datasets with strong class conditioning are easier to clone. It also states that synthetic negatives are computationally cheap once the queue is built, but that **\(N\)** and **\(\rho\)** must be tuned per architecture because excessively many synthetic negatives or too much hardness can hurt. The practical recommendations are correspondingly conservative: tune \(\alpha\) on a held-out set, use **moderate hardness \(N\approx 256\)** and **synthetic ratio \(\rho\approx 0.1\)**, extend pre-training by **\(2\times\text{–}4\times\)** epochs when relying heavily on synthetic data, and monitor representation collapse via inter-class cosine-similarity distributions.

## 4. Syn2Co as a protocol for fluorographdiyne nanosheet synthesis

In the surface-science literature, Syn2Co denotes a selective on-surface 2D covalent polymerization protocol for synthesizing **single-layered fluorographdiyne nanosheets** on **Au(111)** through the combination of **cobalt catalysis** and **coronene templating** [2606.00495]. The target material belongs to the broader family of graphdiyne derivatives with **sp-sp\(^2\)** hybridized skeletons.

The reagents and setup are specified in procedural detail. The precursor is **1,3,5-tris(chloroethynyl)-2,4,6-trifluorobenzene (tFtCEB)**, degassed at **333 K**. The template is **coronene**, which sublimes at **\(\sim 373\ \mathrm{K}\)**. The catalyst is **Co metal** evaporated to **0.05–0.10 ML coverage** using an **e-beam evaporator**. The substrate is an **Au(111) single crystal**, cleaned by **Ar\(^+\) sputtering/annealing under UHV \((p<2\times 10^{-10}\ \mathrm{mbar})\)**, with **base pressure during depositions \(<5\times 10^{-10}\ \mathrm{mbar}\)**.

The synthesis sequence is reported stepwise. First, **0.7 ML tFtCEB** is deposited onto Au(111) held at **250 K**, forming **self-assembled alkynyl–Au–alkynyl dimers**. Annealing to **473 K for 5 min** yields a **large-domain honeycomb sp-MON network** with **Csp–Au–Csp linkages**. Coronene may then be deposited as an optional step at **0.1 ML and 300 K**, producing **coronene-embedded sp-MON**. Next, **0.05–0.10 ML Co** is deposited at **250 K** onto the **(coronene-embedded) sp-MON**. A further anneal to **473 K for 5 min** produces **partial demetallization intermediates**, stated as **\(\sim 50\%\) C–C coupled**. Final annealing to **553 K for 5 min** yields **large-domain fluorographdiyne nanosheets**.

The reported outcome metrics are explicit: **coupling efficiency \(>95\%\)** for the **Csp–Au \(\to\) Csp\(^2\)–Csp\(^2\)** conversion, **hexagon-formation selectivity of \(72.7\%\) without coronene and \(92.3\%\) with coronene**, and **typical covalent domains up to \(60\times 60\ \mathrm{nm}^2\)**. These data place the protocol within the larger effort to obtain large-domain, regular, single-layered graphdiyne-type sheets on surfaces, a problem the paper identifies as a longstanding synthetic challenge.

## 5. Cobalt activation, coronene templating, and reaction energetics

The mechanistic role of cobalt is formulated as **\(d\)–\(\pi\) coordination** between Co and the alkynyl \(\pi\) system:
\[
H_{\text{d–}\pi}
=
\sum_{i\in d,\;j\in \pi}
V_{ij}\,d_i^\dagger\,\pi_j+\mathrm{h.c.},
\]
with \(V_{ij}=\langle d_i|\hat H|\pi_j\rangle\) quantifying **Co(d)–C\(\equiv\)C(\(\pi\)) coupling** [2606.00495]. The reported interpretation is that strong \(d\)–\(\pi\) coupling transforms a robust **Csp–Au** bond into a weaker **Csp\(^2\)–Au** bond, facilitating demetallization and C–C coupling.

The bond-conversion step is written as
\[
\mathrm{R\!-\!C\equiv C\!-\!Au}
+\mathrm{Co}
\xrightarrow{\Delta E=-3.14\,\mathrm{eV}}
\mathrm{R\!-\!C\!=\!C\!-\!Au\!-\!Co}.
\]
In the DFT-optimized structures, the **sp-hybridized C–C(Au) bond length in IS2 is \(1.255\ \mathrm{\AA}\)**, whereas the **Co-activated (sp\(^2\)) C–C(Au) bond length in IS3 is \(1.342\ \mathrm{\AA}\)**. The uncatalyzed surface-stabilized pathway is reported as
\[
\mathrm{Csp\!-\!Au\!-\!Csp}
\xrightarrow{E_a=2.04\,\mathrm{eV}}
\mathrm{Csp\!-\!Csp},
\]
while the Co-catalyzed route is
\[
\mathrm{IS2+2Co}
\xrightarrow{\Delta E=-3.14\,\mathrm{eV}}
\mathrm{IS3}
\xrightarrow{E_a=1.27\,\mathrm{eV}}
\mathrm{FS3+2Co}.
\]
The associated reaction-coordinate data at \(T=0\ \mathrm{K}\) are \(E_{\mathrm{rel}}=0.00\to -3.14\to -1.87\ \mathrm{eV}\) for the catalyzed route, compared with **\(E_{\mathrm{rel}}=-1.24\ \mathrm{eV}\)** for the uncatalyzed pathway. The paper summarizes this as cobalt **lowering the Csp–Au \(\to\) Csp\(^2\)–Au barrier by \(\sim 0.77\ \mathrm{eV}\)**.

Coronene acts as a template that matches the fluorographdiyne lattice and forms **up to six C–H\(\cdots\)F hydrogen bonds per coronene**, with **\(E_{\mathrm{H\!-\!bond}}\approx 0.06\text{–}0.10\ \mathrm{eV}\times 6 \simeq 0.46\ \mathrm{eV}\)** total. The paper’s mechanistic picture states that coronene “locks” partial oligomer radicals via noncovalent anchors, thereby reducing lateral diffusion and rotation. This favors **six-membered motifs** and suppresses kinetically trapped **5/7-MR defects**. The same section reports **0.2–0.4 eV per oligomer stabilization** in the summary statement. Domain-size distributions also shift toward larger domains with coronene: without coronene, the highest fractions lie in the **82–152** and **152–252 \(\mathrm{nm}^2\)** bins; with coronene, weight shifts into the **252–352**, **352–452**, and larger bins.

## 6. Atomic-scale characterization and relation to adjacent terminology

The protocol is supported by atomically resolved microscopy and spectroscopy [2606.00495]. In STM, the **sp-MON honeycomb cell constant is \(0.93\ \mathrm{nm}\)**, and the final fluorographdiyne lattice retains a **hexagonal cell with \(0.93\ \mathrm{nm}\) periodicity**. In **nc-AFM with a CO tip**, the **C\(\equiv\)C\(\cdots\)C\(\equiv\)C diacetylene linkages** appear as **bright rods**. Bond lengths extracted from AFM contrast are given as **\(1.20\pm 0.03\ \mathrm{\AA}\)** for the **C\(\equiv\)C triple bond** and **\(1.34\pm 0.02\ \mathrm{\AA}\)** for a bond with **double character**.

The spectroscopy is similarly specific. At **5 K**, the **fluorographdiyne valence-band resonance** appears at **\(-0.90\ \mathrm{V}\)** and the **conduction-band resonance** at **\(2.15\ \mathrm{V}\)**, corresponding to a measured bandgap of **\(\sim 3.05\ \mathrm{eV}\)**. For coronene, the **HOMO** is at **\(-1.30\ \mathrm{V}\)** and the **LUMO** at **\(2.40\ \mathrm{V}\)**, giving a bandgap of **\(\sim 3.70\ \mathrm{eV}\)**. The spatial maps localize the \(-0.90\ \mathrm{V}\) density on **diacetylene linkages** and the \(+2.15\ \mathrm{V}\) density on **aromatic rings**; the coronene HOMO/LUMO maps match free-molecule LDOS with a **six-lobed flower** HOMO and **ring-shaped** LUMO.

Set against the vision usage, these measurements underscore that the two Syn2Co usages are unrelated except for the acronym. In the vision framework, “synthetic” refers to **diffusion-generated images** and **synthetic feature-space negatives** [2509.02029]. In the surface-science protocol, Syn2Co refers to **Co \(d\)–\(\pi\) activation** and **coronene-regulated ring formation** [2606.00495]. A plausible implication is that “Syn2Co” should be treated as a context-sensitive label rather than as a stable cross-disciplinary term. The nearby acronym **SyCo**, introduced as **Synthetic Coordinate Embedding** for molecular graph generation in latent Euclidean space, reinforces this point: it is adjacent in spelling but methodologically distinct [2406.10513].

Source: https://www.emergentmind.com/topics/syn2co