标签: VCSEL

随着苹果带火了 3D 感测成像技术，作为技术核心的 VCSEL 部分，大家对它的关注度也是水涨船高。 VCSEL 到底有何魅力，能让市场趋之若鹜？也许通过这篇文章，我们可以管中窥豹。 一、 VCSSEL 在光射领域的优势 VCSEL主要有两种，一种用在光通讯领域，波长主要为850 nm，另一种用在消费电子领域，波长主要为940 nm。苹果带火的VCSEL，是消费电子的3D成像领域，主要用于人脸识别，一般常用3D结构光方案和ToF方案。 而用于光通信系统和 3D成像模块的红外光源可以分为三大类：1、发光二极管LED；2、边发射激光器EEL；3、垂直腔面发射激光器VCSEL。 那为什么苹果最后选择了 VCSEL？原因如下：用红外LED作为光源，虽然成本低，但LED没有谐振腔，光束发散，必须输入更多的功率以克服损失，所以其功耗较高。此外，LED不能快速调制，限制了其分辨率。相比于红外LED和EEL，VCSEL的出射光更集中、光斑更对称，同时在温度漂移和腔面反射率上也更占优势。 为什么在激光雷达应用中， VCSEL 也收到如此大的青睐呢？ 激光光源是车载激光雷达的核心器件之一。目前常见的几种光源主要包括边发射激光器 (EEL)、垂直腔面发射激光器(VCSEL)、固体激光器以及光纤激光器等。 激光光源的选择需综合考虑实际应用环境、激光雷达技术方案、性能需求以及成本需求，需要多种类型的激光光源来适应不同的道路环境。 边发射激光器 (EEL)边缘发射激光器具有高功率密度和高脉冲峰值功率，非常适合使用APD探测器的激光雷达系统。VCSEL 具有很高的吸引力，因为它有可能将2D发射器阵列和2D SPAD探测器阵列组合成一个没有运动部件的激光雷达系统。光纤激光器输出功率高、光束质量好、速度快，是高性能系统的理想选择，但其复杂性显著增加。 综上，各波长各有优劣，需结合激光雷达的系统设计从而选择合适的波长。根据不同的驾驶环境需要，边发射型半导体激光器， VCSEL激光器，固体激光器，光纤激光器都有各自的用武之地。 二、 陶瓷基板在 VCSEL 芯片的优势 VCSEL 运行时会产生较大热量。一个是热量需要通过基板及时散发出去；其次 VCSEL 芯片功率密度很高，需要考虑芯片和基板热膨胀失配导致的应力问题。因此，实现高效散热、热电分离及热膨胀系数匹配成为 VCSEL 元件封装基板选择的重要考量，陶瓷基板具备了高导热、高绝缘、高线路精准度、高表面平整度及热膨胀系数与芯片匹配等诸多特性，在 3D 感测、人脸识别、移动照明等各领域中迅速占据了重要地位。 斯利通公司通过选用氧化铝、氮化铝、氮化硅、碳化硅、氧化锆、 ZAT 、蓝宝石等不同材料的陶瓷基板进行加工，不仅具有陶瓷线路板的特性，还可以按照客户的实际要求进行定制。比如客户对电频信号损耗要求很高的产品需要用到介电常数更低的材料二氧化硅线路板。 VCSEL 的固态激光雷达，为汽车领域应用激光雷达。 VCSE 芯片功率转化效率低，这就意味着散热肯定有问题，面临热电分离的难题，而陶瓷基板就是为解决热电分离诞生的。从拆解图片来看， VCSEL 芯片安装在一块氮化铝原料的 DPC 陶瓷基板上，氮化铝基板又贴装于一个 HTCC 陶瓷基座底部。

众所周知，激光雷达跟传感器的结构相似，是由发射器，接收器，数据处理器三部分构成。其中发射器是激光雷达的重要组成部分。根据近几年全球数据显示，半导体激光器大火的要数VCSEL激光发射器。 VCSEL激光器全称为垂直腔面发射激光器，是半导体激光器的一种，当前以砷化镓半导体为基础材料的VCSEL居多，发射波长主要为近红外波段。顾名思义，VCSEL是一种垂直于衬底面射出激光的激光器，有区别于传统的边发射半导体激光器。其可以在衬底上多个方向上排列多个激光器，从而形成并行光源或者面阵光源。 自2017年VCSEL成功应用于Iphon手机人脸识别模组，开始大规模受到关注。同时由于功率提升在激光雷达，安防照明等中长距领域也开始得到应用。 相比固体激光器、光纤激光器等技术，VCSEL半导体激光器的电光转换效率较高，可达40%~60%，即便如此其工作时仍会产生大量的热，如散热效果不佳，则会造成芯片温度升高——直接影响半导体激光器输出功率、阈值电流密度、电光转化效率、微分量子效率、偏振度等性能，并导致半导体激光器寿命和可靠性的下降，甚至会损毁芯片。良好的散热是保障半导体激光器功率和光束质量关键因素之一。 因此高功率 半导体激光器封装 对过渡热沉的要求主要体现在低热阻与低热失配两方面——过渡热沉热导率越高越能有效降低激光器热阻；与此同时还要考虑芯片与热沉热膨胀系数匹配，选择合适的烧结焊料，减少热失配，进而保证/提升激光器输出特性。 激光器的热阻与热导率成反比关系，热沉材料的热导率越高，越可有效降低器件热阻。 芯片与过渡热沉的热膨胀系数失配产生热应力，热应力会影响半导体激光器输出功率、光谱宽度、可靠性等，因此需选用与激光器芯片热膨胀系数匹配的热沉材料。 考虑到以上因素，氮化铝、氧化铝、金刚石等热沉材料应运而生，他们的应用可提升散热能力，减少热阻，提高激光器输出功率，延长激光器寿命。 VCSEL封装是需要把透镜架设到芯片上方，即基板是需要做成三维腔室，而斯利通陶瓷电路板不仅可以做出平面电路板更是可以做出三维电路板——围坝产品，其材质均为无机陶瓷材料，热膨胀系数匹配，制备过程中不会出现脱层、翘曲等现象。 VCSEL的功率密度很高， 芯片与过渡热沉的热膨胀系数失配产生热应力，热应力会影响半导体激光器输出功率、光谱宽度、可靠性等，因此需选用与激光器芯片热膨胀系数匹配的热沉材料。芯片材料为砷化镓，热膨胀系数为 4.5×10-6/K，氮化铝热沉热膨胀系数为 4.1~4.5×10-6/K；金刚石1.0~4.5×10-6/K。 陶瓷基板具有与VCSEL高匹配的热膨胀系数，从而解决则芯片和基板热膨胀失配导致的应力问题。 部分应用领域过度热沉需用到金刚石，其热导率可达1000~2200W/（K·m），相比于热导率为 140~230W/（K·m）的氮化铝过渡热沉，金刚石热沉可显著提高激光器散热效果；但相比于金刚石热沉，使用氮化铝热沉封装芯片的热失配度更低（基于热膨胀参数等数据）。因此采用金刚石热沉作为过渡热沉封装激光器时，对焊料有一定要求，一定要选择合适的焊料以减小热失配引入的热应力。所以目前市场上过度热沉主要还是以氮化铝为主。但是相信在配套焊料不但提高升级的过程中， 金刚石 会和氮化铝一样作为过度热沉一定会成为未来的趋势。

激光雷达的组成部分 由发射系统、接收系统 、信息处理等部分组成. 发射系统是各种形式的激光器，如二氧化碳激光器、掺钕钇铝石榴石激光器、半导体激光器及波长可调谐的固体激光器以及光学扩束单元等组成；接收系统采用望远镜和各种形式的光电探测器，如光电倍增管、半导体光电二极管、雪崩光电二极管、红外和可见光多元探测器件等组合。激光雷达采用脉冲或连续波2种工作方式，探测方法按照探测的原理不同可以分为米散射、瑞利散射、拉曼散射、布里渊散射、荧光、多普勒等激光雷达。 激光雷达的类型及分类 由于当前激光雷达技术方案的分歧点在于扫描方式，所以通常按照扫描方式来分，可以分为：机械旋转激光雷达，混合半固态激光雷达和全固态激光雷达（Flash型和相控阵）。 自动驾驶汽车激光雷达的作用是什么 激光雷达在自动驾驶中的作用，主要是3D/4D环境感知，探测车辆行驶过程中的路况和障碍物，把数据和信号传递给自动驾驶的大脑，再做出相应的驾驶动作。激光雷达可以说是自动驾驶中无形的眼睛. 激光雷达和氮化铝陶瓷基板 激光雷达的激光发生器--VCSEL激光器全名为垂直共振腔表面放射激光器，简称面射型激光器。它以砷化镓半导体材料为基础研制，是一种半导体激光器。VCSEL的固态激光雷达具有更高的可靠性、稳定性并尺寸小型化，为汽车领域大规模应用激光雷达奠定了基础。 VCSEL芯片功率转化效率较低，这就意味着散热肯定有问题，面临热电分离的难题，而陶瓷基板就是为解决热电分离诞生的。根据有关拆解图片来看，VCSEL芯片安装在一块氮化铝原料的DPC陶瓷基板上，氮化铝基板又贴装于一个陶瓷基座底部。 VCSEL运行时会产生较大热量。其一，一个是热量需要通过基板及时散发出去；其次，VCSEL芯片功率密度很高，需要考虑芯片和基板热膨胀失配导致的应力问题。因此，实现高效散热、热电分离及热膨胀系数匹配成为VCSEL元件封装基板选择的重要考量。 一般情况下，半导体激光器的发光波长随温度变化为0.2-0.3nm/℃，光谱宽度随之增加，影响颜色鲜艳度。另外，当正向电流流经pn结，发热性损耗使结区产生温升，在室温附近，温度每升高1℃，半导体激光器的发光强度会相应地减少1%左右，激光器时刻保持色纯度与发光强度非常重要，以往多采用减少其驱动电流的办法，降低结温，多数半导体激光器的驱动电流限制在20mA左右。但是，半导体激光器的光输出会随电流的增大而增加，很多功率型半导体激光器的驱动电流可以达到70mA、100mA甚至1A级，需要改进封装结构才能保证激光器的寿命，全新的半导体激光器封装设计理念采用低热阻封装结构及技术，改善热特性。 这就要求选择与芯片材料匹配的热膨胀系数接近的且有高热导的的封装基板--氮化铝的DPC陶瓷基板。请看下表： 氮化铝的热膨胀系数不仅与砷化镓半导体材料和非常接近，且具备180W/M*K)的热导率。 直接镀铜陶瓷基板DPC陶瓷基板极大地满足了VCSEL元件的这种封装要求。由于DPC陶瓷基板具备了高导热、高绝缘、高线路精准度、高表面平整度及热膨胀系数与芯片匹配等诸多特性，在高功率VCSEL元件封装中占有重要地位。 由于VCSEL的结构是垂直结构，斯利通DPC陶瓷电路板具有独特的高解析度、高平整度及高可靠垂直互联等技术优势更适用于其垂直共晶焊接。 陶瓷本身的稳定性确保传感器信号不会失真； 陶瓷基板 与芯片的热膨胀系数匹配，使得产品更加可靠，即使在汽车高温，高震动，含腐蚀性的环境下仍然可以保正信号的高效，灵敏，准确。

We tend to picture lasers, regardless of power rating, as highly focused, coherent light sources. After all, one of the virtues of the laser is that its beam doesn't spread, so it can be used for targeted illumination (such as scanner or rangefinder), or high-intensity localized heating (cutting or welding metals), to cite just a few of their thousands of applications. (Historical side note: when the laser was first demonstrated, one pundit wagged "it was a solution looking for a problem to solve," and we know how that quip turned out!)   But using lasers for area heating seems to be a contrary to their virtues. After all you can heat with IR lamps, microwaves, heated air, electric coils, induction coils, or gas-fired burners, to cite just a few possibilities. Why would you go to the complications of using lasers unless you had no alternative?   That's why I was surprised when I saw the story on the benefits of "photonic" heating in Laser Focus World, "High-power VCSEL arrays make ideal industrial heating systems." By setting up an array of hundreds of vertical-cavity surface-emitting lasers (shown below), you not only obviously get a different source of heat, but you attain some other unique operating advantages that are non-obvious and beneficial. Yes, the author's company (Photonics Aachen, part of Philips Photonics ) makes this system and so he is somewhat biased, but nonetheless, it's worth seeing what he has to say.   A one or two-dimensional array of VCSELs can be used as tightly spaced, easily modulated high-intensity heat source. (Source: Philips Photonics)   In contrast to conventional edge-emitting laser diodes, the collection of vertical-emitting laser diodes (each with a diameter of 30-40 μmeter) can be fabricated in one pass of wafer processing, including test, with about 500 VCSELs per mm 2 of a die. Since each laser emits 1 to 10 mW, a 2 × 2 mm chip array holding 2,000 VCSELs can emit over 20 W of infrared power — that's impressive power density for this technology. These arrays can be connected in series, so arrays of hundreds of watts and even kW have been built. While heat sinking of the die is an issue, it’s a manageable one, the author claims.   All this is impressive, but why bother when you can use standard halogen lamps, for example, to get the IR heating? First, the VCSEL IR brightness is 100 to 1,000 greater than halogens, with a lifetime of over than 10,000 hours, the author says (and I'll have to accept those numbers for now).   But the advantages of VCSEL-based heating go beyond just density and lifetime. The VCSEL array can be switched on and off in milliseconds for precise dosing control, since it does not have the thermal lag of a halogen emitter or similar sources. Also, the VCSEL array is well suited to highly targeted, localized zones, where the material to be heated may not be homogenous, with some areas needing more or less heat or specialized heat-treating patterns are preferred (think of PC boards to be wave soldered and loaded with different-size/mass components). Even if this level of control is not a requirement, the output power and thus heat is tightly focused, so that the entire oven does not have to be heated; only the part of the material that needs treatment. Further, unlike bulbs, the VCSEL's output wavelength does not change as it is dimmed, which means the target material's thermal absorption characteristics are unchanged, a factor in precision situation. Dimming control is relatively easy, as the VCSELs are driven by a controllable DC current.   I don’t have a need to heat anything with VCSELs, of course, nor do I fully understand what downsides of this heating approach. Still, reading about this application reminded me that what we often consider a key attribute of a technology can sometimes be less relevant, while its downside can become a virtue. While we prize lasers for their ability to deliver highly focused beams of photonic power, and use ever-bigger single-source lasers to deliver increasingly powerful punches, some out-of-the-box thinking shows that an aggregation of many small lasers as heat sources can be used to advantage for some applications.   Have you seen other cases where contrary thinking has solved a power problem or turned a thermal weakness into an advantage? Have you ever done this?

We tend to picture lasers, regardless of power rating, as highly focused, coherent light sources. After all, one of the virtues of the laser is that its beam doesn't spread, so it can be used for targeted illumination (such as scanner or rangefinder), or high-intensity localized heating (cutting or welding metals), to cite just a few of their thousands of applications. (Historical side note: when the laser was first demonstrated, one pundit wagged "it was a solution looking for a problem to solve," and we know how that quip turned out!)   But using lasers for area heating seems to be a contrary to their virtues. After all you can heat with IR lamps, microwaves, heated air, electric coils, induction coils, or gas-fired burners, to cite just a few possibilities. Why would you go to the complications of using lasers unless you had no alternative?   That's why I was surprised when I saw the story on the benefits of "photonic" heating in Laser Focus World, "High-power VCSEL arrays make ideal industrial heating systems." By setting up an array of hundreds of vertical-cavity surface-emitting lasers (shown below), you not only obviously get a different source of heat, but you attain some other unique operating advantages that are non-obvious and beneficial. Yes, the author's company (Photonics Aachen, part of Philips Photonics ) makes this system and so he is somewhat biased, but nonetheless, it's worth seeing what he has to say.   A one or two-dimensional array of VCSELs can be used as tightly spaced, easily modulated high-intensity heat source. (Source: Philips Photonics)   In contrast to conventional edge-emitting laser diodes, the collection of vertical-emitting laser diodes (each with a diameter of 30-40 μmeter) can be fabricated in one pass of wafer processing, including test, with about 500 VCSELs per mm 2 of a die. Since each laser emits 1 to 10 mW, a 2 × 2 mm chip array holding 2,000 VCSELs can emit over 20 W of infrared power — that's impressive power density for this technology. These arrays can be connected in series, so arrays of hundreds of watts and even kW have been built. While heat sinking of the die is an issue, it’s a manageable one, the author claims.   All this is impressive, but why bother when you can use standard halogen lamps, for example, to get the IR heating? First, the VCSEL IR brightness is 100 to 1,000 greater than halogens, with a lifetime of over than 10,000 hours, the author says (and I'll have to accept those numbers for now).   But the advantages of VCSEL-based heating go beyond just density and lifetime. The VCSEL array can be switched on and off in milliseconds for precise dosing control, since it does not have the thermal lag of a halogen emitter or similar sources. Also, the VCSEL array is well suited to highly targeted, localized zones, where the material to be heated may not be homogenous, with some areas needing more or less heat or specialized heat-treating patterns are preferred (think of PC boards to be wave soldered and loaded with different-size/mass components). Even if this level of control is not a requirement, the output power and thus heat is tightly focused, so that the entire oven does not have to be heated; only the part of the material that needs treatment. Further, unlike bulbs, the VCSEL's output wavelength does not change as it is dimmed, which means the target material's thermal absorption characteristics are unchanged, a factor in precision situation. Dimming control is relatively easy, as the VCSELs are driven by a controllable DC current.   I don’t have a need to heat anything with VCSELs, of course, nor do I fully understand what downsides of this heating approach. Still, reading about this application reminded me that what we often consider a key attribute of a technology can sometimes be less relevant, while its downside can become a virtue. While we prize lasers for their ability to deliver highly focused beams of photonic power, and use ever-bigger single-source lasers to deliver increasingly powerful punches, some out-of-the-box thinking shows that an aggregation of many small lasers as heat sources can be used to advantage for some applications.   Have you seen other cases where contrary thinking has solved a power problem or turned a thermal weakness into an advantage? Have you ever done this?