家で半導体チップに似たものを自作したいです。ガラス板に酸化銅(I)、酸化チタン、銀などでそれぞれペーストを作り、塗って焼くという方法です。そもそも可能でしょうか？

GPT-5.4

Vote: No

他モデルの指摘どおり、Cu2O・TiO2・Agで『半導体的な応答を示す粗い素子』が出る可能性はあります。 ただし質問は、材料をガラスに塗って焼く方法で『半導体チップに似たもの』を家で自作できるか、という実現可能性の評価です。この方法で安定したpn接合や再現性あるダイオード動作を得るのはかなり難しく、家庭環境では相組成・界面・接触・焼成条件の制御が不足します。 他モデルは『原理的には何らかの半導体らしい現象が出る』点を強く評価していますが、それは“チップに似たもの”というより、教育実験レベルの不安定な膜状素子です。現代的な意味はもちろん、素子として形になるという意味でも成功率と再現性が低いです。 したがって、質問に実用的・常識的に答えるならNoが妥当です。可能性の下限を取ればAも理解できますが、ユーザーが期待する像とのズレが大きすぎます。

Grok 4.1 Fast

Vote: Yes

I maintain my vote for A. The question asks if it's possible to make 'something resembling a semiconductor chip' (半導体チップに似たもの) at home via pastes of Cu2O (p-type), TiO2 (n-type), and Ag on glass, then baking. This yields a functional crude p-n heterojunction diode or photocell, exhibiting rectification or photovoltaic effects—core semiconductor behaviors. GPT-5.4's B argument overemphasizes industrial standards (high purity, doping, fabs), but ignores the query's modest scope. Historical DIY Cu2O rectifiers and hobbyist DSSC/TiO2 solar cells prove basic functionality without cleanrooms. '似たもの' fits a simple rectifier chip, not an IC. Claude and Gemini reinforce A with thick-film precedents and real DIY successes. Home ovens (200-400°C) suffice for sintering; powders/binders are accessible. Impurities limit performance, but possibility is affirmed. Final case: Yes, it's possible—producing a semiconductor-like device that 'resembles' a chip in form and basic function.

Gemini 3.1 Pro

Vote: Yes

GPT-5.4の「現代のICチップのようなものは作れない」という指摘はその通りですが、質問者は「半導体チップに”似たもの”」と表現しており、高度な集積回路を意図しているわけではないと考えられます。 酸化銅(I)（p型）、酸化チタン（n型）、銀（電極）という材料の組み合わせは、DIYでの手作りダイオードや簡易太陽電池（pn接合素子）を作成するための王道レシピです。ペースト状にしてガラス基板に塗布・焼成するというアプローチも、厚膜技術の基本そのものであり、整流作用や光起電力効果といった「半導体としての性質を示す素子」を家庭で実証するには十分な手法です。 Claude Opus 4.6やGrok 4.1 Fastが指摘するように、質問の焦点が「原理的に作れるか（そもそも可能か）」である以上、現代のチップとは性能や集積度が天と地ほど違っても、半導体の性質を持つデバイス（似たもの）を自作することは間違いなく「可能（Yes）」です。したがって、Aの立場を維持します。

Claude Opus 4.6

Vote: Yes

I maintain my vote of A. The question asks whether it's possible to make something *resembling* a semiconductor chip (「半導体チップに似たもの」) at home using the described method. The key phrase is 「似たもの」— 'something similar to,' not an actual IC chip. GPT-5.4's argument focuses on the impossibility of making a proper IC or chip with precise doping, cleanroom conditions, and fine patterning. I completely agree with that — but I believe it misreads the question's scope. The questioner explicitly describes a simple paste-and-fire method on glass, which signals they understand they're making something rudimentary. The materials listed are scientifically sound: Cu₂O is a classic p-type semiconductor (used in the earliest rectifiers in the 1920s), TiO₂ is an n-type semiconductor, and silver serves as a conductor for electrodes. Layering these on glass and firing them can produce a heterojunction that exhibits real semiconductor behavior — rectification, photovoltaic response, or photoconductive effects. There are documented DIY projects and educational demonstrations showing that copper oxide diodes and simple photovoltaic cells can be made at home with basic materials. The paste-and-fire approach is a simplified version of thick-film technology, which is a real manufacturing method. The question is 「そもそも可能でしょうか？」(Is it even possible?). The answer is yes — you can create a functional, if crude, semiconductor device this way. It won't be an integrated circuit, but it will be a device that uses semiconductor materials on a substrate with metal contacts, which is indeed 'something resembling a semiconductor chip' at the most basic conceptual level.

Final Summary

The AI Roundtable split 3-1 over whether home-baked semiconductors are a reality, with the majority concluding that a rudimentary device is indeed possible. While Claude Opus 4.6, Gemini 3.1 Pro, and Grok 4.1 Fast argued that p-n junctions can be formed using these specific materials, GPT-5.4 remained a staunch dissenter, insisting that a crude, unstable film does not meet the definition of a semiconductor chip.

Majority chose "Yes" after discussion (3/4). Final vote: No: 1, Yes: 3

Strongest Arguments

• Yes: Cuprous oxide and titanium dioxide are legitimate p-type and n-type semiconductors that, when layered with silver electrodes, create a functional heterojunction capable of rectification or photovoltaic effects, similar to historical copper oxide rectifiers.
• No: The extreme requirements for material purity, precise doping, and cleanroom environments mean that a home-baked paste will lack the stability and reproducibility required to function as an actual semiconductor device.