文件类型：PPT/Microsoft Powerpoint 文件大小：字节
Data Analysis
数据分析
Turning signals into answers
从信号中获取信息
Making Sense of the Numbers
弄清数字的含义
Signal Conditioning makes nice voltages
信号预处理 消除各种噪声
Data Normalization makes spectra
数据标准化 得到光谱
Spectral Processing makes nice spectra
光谱处理
Spectral Modeling makes nice answers
光谱建模 实际应用
Signal Conditioning
信号预处理
Dark Spectra Subtraction 噪声光谱消除
Electrical Dark Correction 电子噪声校正
Scans to Average 扫描平均次数
Box Car smoothing 平滑度系数
Linearity Correction 线性度校正
Stray Light Correction 光漂移校正
Dark Spectra Subtraction
光谱噪声消除
Dark spectra is stored in dark data array
点此按钮将光谱噪声存储在内存中
Dark spectra is subtracted for on-screen display
点此按钮在屏幕显示上减去噪声光谱
Dark is subtracted as part of algorithm for A,T,I
在吸收,透射和发射测量中,噪声光谱的消除是计算中的一环
Dark Spectra is the dark current plus the amplifier offset (electrical dark) plus an offset that is unique for each pixel. The offset is called "Fixed Pattern" noise.
噪声光谱=暗电流+放大器偏移(电子噪声)+每个像素的偏移(固定模式噪声)
"Fixed Pattern" noise is removed by subtracting a stored dark spectra
固定模式噪声可以在消除噪声光谱的过程中抵消
Subtracting Stored dark removed 98% of the pattern
噪声光谱的消除过程中抵消了98%的固定模式噪声
When to Subtract Dark
何时需要消除噪声光谱
The dark spectra must be subtracted for any quantitative use of the data
在任何定量分析中都需要消除噪声光谱
Because of the fixed pattern, you must subtract the dark level for each individual pixel (can't use a global dark)
由于固定模式噪声的存在,必须针对每个像素去除噪声
Fixed pattern and dark level increase with longer integration periods, and for cosmetic purposes dark subtract should be used for any experiment where integration is > 500 msec or so
固定模式噪声和光谱噪声随积分时间增加而增加,在积分时间大于500毫秒左右时需要消除噪声光谱
Electrical Dark Correction
电子噪声校正
First 24 pixels are masked, don't respond to light
头24个像素被屏蔽不能接收光
The average of pixels 3-24 for each scan is used to calculate a global "electrical dark correction factor"
第3到24像素的平均值被用于计算电子噪声校正因子
When electrical dark correction is turned on, the global factor is subtracted from each pixel value
当打开电子噪声校正时每个像素都减去这个因子值
The correction factor is filtered, by using a moving average of the last 10 scans.
校正因子采用最后10次扫描的平均值
When to use Correct for Electrical dark
何时采用电子噪声校正
The correction does not directly affect data quality, it is primarily cosmetic
该校正并不实际影响数据质量,只影响显示效果
You must take sample, reference and dark spectra the same way, either with the electrical dark correction on or off.
必须采用同样的方式来采样信号,参考值及噪声光谱值,无论电子噪声校正是否开启
It is useful for correcting the drift in amplifier electronics caused by temperature (as all changes in amplifier offset will be subtracted out).
在校正由于温度引起的电子放大漂移时很有用,因为放大电路偏移的影响可以被抵消
Scans to Average
扫描平均次数
Signal to noise improves (and random noise is reduced) by averaging more data.
提高信噪比(减小随机噪声)
The number of scans to average (n) is set in software
扫描平均次数在软件中设置
The improvement is = (n)1/2
信噪比提高率=(n)1/2
How many scans should I average
如何设置扫描平均次数
More is better
越多越好
The trade-off is time. The software only updates the graph after the scans have been averaged.
扫描平均次数越多需要时间越长.
If the sample is dynamic, the appropriate averaging is dictated by how fast the sample changes
如果采样数据是动态的,合适的平均次数取决于样本的变化速度
For static samples, the patient of the experimenter sets the limit (or other long term drift issues such as the light source, temperature, etc.)
对于静态样本,需要考虑的是在长时间的温度和光源漂移问题
Box Car Averaging (smoothing)
平滑度
The box car setting (n) controls averaging across the pixels in a single scan.
平滑度用来控制单次扫描的按像素的平均
Each pixel is reported as the avaerage of that pixel and its n nearest neighbors (higher and lower)
每个像素的值采用左右两边n个像素值的平均
S:N improves with (2n+1)1/2
信噪比提高率为 (2n+1)1/2
Boxcar Smoothing – Trade-off
平滑度影响的因素
Box car averaging alters the spectral shape.
改变了光谱的形状
Sharp features are affected more than broad features.
尖峰的改变比宽峰改变大
The best setting is to match the resolution of the slit.
最佳设置值要匹配狭缝的分辨率
Should you uses smoothing It depends on your application.
是否需要平滑根据具体应用来定
Detector linearity
探测器线性度
In ideal sensor 理想情况下Vout = k Iin
In image sensors, Vout deviates from linearity.
Vout会有线性偏离
We test linearity by keeping constant Iin, and varying the Vout using the integration time:在检测线性度时保持Iin不变,通过改变积分时间改变Vout
Vout/Dt = kIin
In software we correct for linearity.
在软件中作线性度校正
Electrical Dark
电子噪声
Amplifier set point, controlled by a potentiometer on the circuit card.
放大器基准点受电路板上电位计的控制
First 14 pixels are "masked" and do not respond to light
头14个像素被屏蔽,对光不响应
Software Electrical Dark Correction takes a dynamic average of first 14 pixels and subtracts this value in real time.
电子噪声软件校正采用头14个像素的动态平均值
The correction factor is filtered by averaging the values from the last 12 scans
校正因子采用最后12次扫描的平均值
Deep well detector has highest response at low signals, and drops off by as much as 2% at high signals. The non-linearity is the same for all pixels
井深大的探测器对弱信号的响应很高,在强信号时会降低约2%
The S2000 detector has a peak response at mid level signals, and drops by 4% at high and low signals.
S2000系列的探测器峰值响应在中间,最大最小值差异达4%
Non Linearity Correction
非线性校正
Correction parameters are factory determined and loaded on EPROM (for USB devices)
校正系数出厂时存在EPROM中(对USB设备)
Correction turned on/off in the Spectrometer Configuration screen
校正选项可以在光谱仪配置屏幕上开启和关掉
Each scan is corrected on the fly.
每次扫描都需要随时校正
When to Use Correct for Linearity
何时使用线性度校正
All the time! There isn't any good reason to not use it
任何时候
Errors can be large if for example the sample is at 2000 counts and the reference at 4000 counts:
比如在采样光谱为2000counts而参考光谱为4000counts时,误差可能较大
Abs error = 0.013
%T error = 1.5%
Errors appear to be unpredictable, if you change conditions and signal levels
如果测试条件和信号大小改变,误差可能不可预期
Stray Light Correction
光漂移校正
Stray light is a global value subtracted from each pixel
光漂移是一个从每个像素都要去除的全局值
The global value is computed for each scan by looking at the total signal across the array and multiplying the % of full signal x the Correction Constant
Stray Light affects linearity of Measurements
光漂移对测量线性度的影响
Error increases as absorbance increase (or %T decreases
吸收率增加时误差增大
Stray light level depends on experiment details, light source, sample characteristics, sampling optics
光漂移水平取决于具体实验细节
User can evaluate stray light using a high concentration, highly absorbing sample of the material being analyzed
对光漂移的评估可以采用高浓度,高吸收率的样品来分析
Stray Light Experiment: Fluorescent light, USB2000-VIS/NIR. Blocking provided by two OG550 filters (OD>7). Stray light was about 0.4 counts. Stray light software correction was set to 17. This gave closest fit to zero stray light. When the correction is "on" stray light was reduced to below the noise level of the measurement
光漂移实验:荧光灯,USB4000-VIS/NIR. 两个OG550滤光片(OD>7).光漂移大约0.4 counts.光漂移软件校正设置在17,这样最接近零漂移.当校正选项开启时光漂移减少到噪声检测水平以下.
Data Normalization 数据标准化
Spectrometer Signal Comes From:
Collection Optics Efficiency
Optical Fiber Attenuation
Optical Bench Efficiency
Fiber Mode Structure
Detector Response & linearity
Detector Interference Pattern
Pixel Fixed Pattern Noise
Amplifier electrical zero
and of course… the sample spectra
光谱信号受下列因素影响:
光采集效率
光纤响应曲线
光学平台的效率
光纤模式
探测器响应及线性度
探测器干扰
像素固定模式噪声
放大电路电子零点
采样光谱
Data Normalization 数据标准化
S(i) - D(i)
R(i) - D(i)
T(i) =
S(i) - D(i)
R(i) - D(i)
A(i) = -log
I(i) = B(i)
S(i) - D(i)
R(i) - D(i)
Percent of Reference Spectra
参考光谱的比例
Transmission, Reflection
透射,反射
Optical Density or Absorbance
光强或吸收 -log of T
Relative Irradiance
相对辐射量
Energy 能量
Choice of Reference for Transmission
透射参考值的选择
Choice of Reference for Reflection
反射测量的参考值选择
Choice of Reference for Energy
光强测量参考值
Normalized Data
数据标准化
T
I
A
Who Needs Data Normalization
谁需要数据标准化
Everybody! 任何人
Scope Mode data can't be compared with any other spectrometer
Scope模式的数据不能和其它任何光谱仪数据相比
Scope Mode is not quantitative if used at more than one wavelength (or pixel)
Scope模式在超过一个波长(像素)时不能量化
Spectra are distorted in shape 光谱形状被扭曲
Except 除非
If you know all of that and you calibrate scope mode to some external set of standards, for example when setting up a fluorescence signal vs concentration curve.
在Scope模式下为所有标准校正,例如荧光信号相对于浓度曲线的校正.
Spectral Processing
光谱处理
Correct Sample channel for drift observed in reference channel
为参考通道的漂移校正样品通道
Internal Standards (wavelength ratios)
内部标准(波长比率)
Math applied to whole spectra sets:
应用到整个光谱分析的数学知识
Subtracting a background spectra 去除背景噪声光谱
Dividing by a reference spectra 除以参考光谱值
Derivativation (1st and 2nd derivative of Amplitude vs Wavelength)
求导(振幅对波长的一阶和二阶导数)
Deconvolution & Curve Fitting 去卷积和曲线拟合
Gaussian, Lorentzian 高斯,拉葛朗日插值
FFT, peak integration, interpolation, etc.
FFT,峰值积分,插值
Sample样品光谱
Reference参考光谱
Reference参考光谱
sample样品光谱
For Absorbance对吸收率测量,Scor(l) = Sobs(l) - R(l)
For Transmission对投射率测量,Scor(l) = Sobs(l) / R(l)
Monitor lamp drift
监测灯光的漂移
Monitor lamp & Sample drift
监测灯光和样品的漂移
Reference Monitoring Options
参考值监测选项
Any channel of spectrometer can use any other channel (or itself) as the reference monitor
每个通道的光谱仪可以利用其它通道的光谱仪作为参考光谱监视器
Three Options 三种选择
Single Point 单点
Wavelength by Wavelength逐波长
Integrated band 积分带
Who Needs Reference Monitoring
谁需要监视参考光谱
Any long term experiment when taking a new reference is not an option
任何没有及时获取参考光谱的长时间实验
Any experiment that looks at the difference between two samples
任何在两次测量有差异的实验中
Any experiment that has samples with interfering backgrounds
样品有背景干扰的实验
Trade-offs – Answer is only as good as the reference monitoring data. If its not necessary, don't use it.
非必要,无需用
Wavelength Ratios & Baseline Corrections – Internal references
波长比率和基准线校正-内部参考
Simple Algorithms can be written in the Time Series Analysis Function
样品光谱可以在时间序列分析功能中计算
Full Spectral Math Functions are available in OOIBase32 Platinum
在OOIBase32白金版中有所有关于光谱计算的数学函数
AddConstant
Adds a constant to all points in a spectrum
每个点加上一个常量
AddSpectra
Adds two spectra, point-by-point
逐点相加两个光谱
Average
Calculates the average value of a spectrum
计算光谱的平均值
AverageWindow
Calculates the average value of a spectral window
计算一段光谱的平均值
Boxcar
Boxcar smooths (non-weighted averaging) a spectrum
对一个光谱进行平滑处理
CalculateAbsorbance
Calculates absorbance for a set of spectra
计算一系列光谱的吸收率
CalculateIrradiance
Calculates relative irradiance for a set of spectra
计算一系列光谱的相对辐射强度
CalculateTransmission
Calculates transmission for a set of spectra
计算一系列光谱的投射率
Derivative
Calculates the derivative of a spectrum
计算光谱的导数
DivideConstant
Divides a constant into all points in a spectrum
光谱每个点除以一个常量
DivideSpectra
Divides two spectra, point-by-point
两光谱逐点相除
GetMaxValue
Calculates the maximum value of a spectrum
计算光谱的极大值
GetMinValue
Calculates the minimum value of a spectrum
计算光谱的极小值
Integrate
Integrates an entire spectrum
在整个光谱上积分
IntegrateWindow
Integrates a region of a spectrum
在一段光谱上积分
Interpolate
Interpolates (using cubic splines) a spectrum to evenly-spaced wavelength intervals
在光谱上进行等距插值(三次样条)
Log10Spectrum
Computes the log10 of a complete spectrum
光谱取对数
MultiplyConstant
Multiplies a constant to all points in a spectrum
在整个光谱上乘以一个常数
MultiplySpectra
Multiplies two spectra, point-by-point两个光谱逐点相乘
PolynomialRegression
Computes a polynomial regression on a set of data
计算多项式回归系数
SubtractConstant
Subtracts a constant to all points in a spectrum
在整个光谱上逐点减去一个常数
SubtractSpectra
Subtracts two spectra, point-by-point两个光谱逐点相减
Sum
Sums all points in a spectrum
整个光谱求和
SumWindow
Sums a region of a spectrum
一段光谱求和
Spectral Modeling
光谱建模
Chromaticity (L* a* b*)色度
Provides a scale related to the appearance of the color to a human eye
提供人眼区分颜色的梯度系统
Beer's Law 比尔法则
Used to calculate concentration of an analyte
计算浓度
Principle Component Analysis 主成分分析
Used to analyze samples with more than one analyte
分析多种混合样品
PLS and other Chemometric Tools 最小平方法及其它
Used to analyze more complex mistures of signals
用于分析更多复杂的混合信号
Spectral Modeling
光谱建模
Neural Nets 神经网络
A training approach to build an empirical model
通过训练的方法建立实验模型
Radiometry 幅射线测定
Measures radiant energy
测量辐射能量
Thin film analysis 薄膜分析
Metrology of film thickness from interference spectra
通过干涉光谱分析薄膜厚度
Scattering & particle size analysis 散射和粒子大小分析
Analyzing particle size from scattering spectra
通过散射光谱分析粒子大小
Thanks
谢谢!
数据分析
Turning signals into answers
从信号中获取信息
Making Sense of the Numbers
弄清数字的含义
Signal Conditioning makes nice voltages
信号预处理 消除各种噪声
Data Normalization makes spectra
数据标准化 得到光谱
Spectral Processing makes nice spectra
光谱处理
Spectral Modeling makes nice answers
光谱建模 实际应用
Signal Conditioning
信号预处理
Dark Spectra Subtraction 噪声光谱消除
Electrical Dark Correction 电子噪声校正
Scans to Average 扫描平均次数
Box Car smoothing 平滑度系数
Linearity Correction 线性度校正
Stray Light Correction 光漂移校正
Dark Spectra Subtraction
光谱噪声消除
Dark spectra is stored in dark data array
点此按钮将光谱噪声存储在内存中
Dark spectra is subtracted for on-screen display
点此按钮在屏幕显示上减去噪声光谱
Dark is subtracted as part of algorithm for A,T,I
在吸收,透射和发射测量中,噪声光谱的消除是计算中的一环
Dark Spectra is the dark current plus the amplifier offset (electrical dark) plus an offset that is unique for each pixel. The offset is called "Fixed Pattern" noise.
噪声光谱=暗电流+放大器偏移(电子噪声)+每个像素的偏移(固定模式噪声)
"Fixed Pattern" noise is removed by subtracting a stored dark spectra
固定模式噪声可以在消除噪声光谱的过程中抵消
Subtracting Stored dark removed 98% of the pattern
噪声光谱的消除过程中抵消了98%的固定模式噪声
When to Subtract Dark
何时需要消除噪声光谱
The dark spectra must be subtracted for any quantitative use of the data
在任何定量分析中都需要消除噪声光谱
Because of the fixed pattern, you must subtract the dark level for each individual pixel (can't use a global dark)
由于固定模式噪声的存在,必须针对每个像素去除噪声
Fixed pattern and dark level increase with longer integration periods, and for cosmetic purposes dark subtract should be used for any experiment where integration is > 500 msec or so
固定模式噪声和光谱噪声随积分时间增加而增加,在积分时间大于500毫秒左右时需要消除噪声光谱
Electrical Dark Correction
电子噪声校正
First 24 pixels are masked, don't respond to light
头24个像素被屏蔽不能接收光
The average of pixels 3-24 for each scan is used to calculate a global "electrical dark correction factor"
第3到24像素的平均值被用于计算电子噪声校正因子
When electrical dark correction is turned on, the global factor is subtracted from each pixel value
当打开电子噪声校正时每个像素都减去这个因子值
The correction factor is filtered, by using a moving average of the last 10 scans.
校正因子采用最后10次扫描的平均值
When to use Correct for Electrical dark
何时采用电子噪声校正
The correction does not directly affect data quality, it is primarily cosmetic
该校正并不实际影响数据质量,只影响显示效果
You must take sample, reference and dark spectra the same way, either with the electrical dark correction on or off.
必须采用同样的方式来采样信号,参考值及噪声光谱值,无论电子噪声校正是否开启
It is useful for correcting the drift in amplifier electronics caused by temperature (as all changes in amplifier offset will be subtracted out).
在校正由于温度引起的电子放大漂移时很有用,因为放大电路偏移的影响可以被抵消
Scans to Average
扫描平均次数
Signal to noise improves (and random noise is reduced) by averaging more data.
提高信噪比(减小随机噪声)
The number of scans to average (n) is set in software
扫描平均次数在软件中设置
The improvement is = (n)1/2
信噪比提高率=(n)1/2
How many scans should I average
如何设置扫描平均次数
More is better
越多越好
The trade-off is time. The software only updates the graph after the scans have been averaged.
扫描平均次数越多需要时间越长.
If the sample is dynamic, the appropriate averaging is dictated by how fast the sample changes
如果采样数据是动态的,合适的平均次数取决于样本的变化速度
For static samples, the patient of the experimenter sets the limit (or other long term drift issues such as the light source, temperature, etc.)
对于静态样本,需要考虑的是在长时间的温度和光源漂移问题
Box Car Averaging (smoothing)
平滑度
The box car setting (n) controls averaging across the pixels in a single scan.
平滑度用来控制单次扫描的按像素的平均
Each pixel is reported as the avaerage of that pixel and its n nearest neighbors (higher and lower)
每个像素的值采用左右两边n个像素值的平均
S:N improves with (2n+1)1/2
信噪比提高率为 (2n+1)1/2
Boxcar Smoothing – Trade-off
平滑度影响的因素
Box car averaging alters the spectral shape.
改变了光谱的形状
Sharp features are affected more than broad features.
尖峰的改变比宽峰改变大
The best setting is to match the resolution of the slit.
最佳设置值要匹配狭缝的分辨率
Should you uses smoothing It depends on your application.
是否需要平滑根据具体应用来定
Detector linearity
探测器线性度
In ideal sensor 理想情况下Vout = k Iin
In image sensors, Vout deviates from linearity.
Vout会有线性偏离
We test linearity by keeping constant Iin, and varying the Vout using the integration time:在检测线性度时保持Iin不变,通过改变积分时间改变Vout
Vout/Dt = kIin
In software we correct for linearity.
在软件中作线性度校正
Electrical Dark
电子噪声
Amplifier set point, controlled by a potentiometer on the circuit card.
放大器基准点受电路板上电位计的控制
First 14 pixels are "masked" and do not respond to light
头14个像素被屏蔽,对光不响应
Software Electrical Dark Correction takes a dynamic average of first 14 pixels and subtracts this value in real time.
电子噪声软件校正采用头14个像素的动态平均值
The correction factor is filtered by averaging the values from the last 12 scans
校正因子采用最后12次扫描的平均值
Deep well detector has highest response at low signals, and drops off by as much as 2% at high signals. The non-linearity is the same for all pixels
井深大的探测器对弱信号的响应很高,在强信号时会降低约2%
The S2000 detector has a peak response at mid level signals, and drops by 4% at high and low signals.
S2000系列的探测器峰值响应在中间,最大最小值差异达4%
Non Linearity Correction
非线性校正
Correction parameters are factory determined and loaded on EPROM (for USB devices)
校正系数出厂时存在EPROM中(对USB设备)
Correction turned on/off in the Spectrometer Configuration screen
校正选项可以在光谱仪配置屏幕上开启和关掉
Each scan is corrected on the fly.
每次扫描都需要随时校正
When to Use Correct for Linearity
何时使用线性度校正
All the time! There isn't any good reason to not use it
任何时候
Errors can be large if for example the sample is at 2000 counts and the reference at 4000 counts:
比如在采样光谱为2000counts而参考光谱为4000counts时,误差可能较大
Abs error = 0.013
%T error = 1.5%
Errors appear to be unpredictable, if you change conditions and signal levels
如果测试条件和信号大小改变,误差可能不可预期
Stray Light Correction
光漂移校正
Stray light is a global value subtracted from each pixel
光漂移是一个从每个像素都要去除的全局值
The global value is computed for each scan by looking at the total signal across the array and multiplying the % of full signal x the Correction Constant
Stray Light affects linearity of Measurements
光漂移对测量线性度的影响
Error increases as absorbance increase (or %T decreases
吸收率增加时误差增大
Stray light level depends on experiment details, light source, sample characteristics, sampling optics
光漂移水平取决于具体实验细节
User can evaluate stray light using a high concentration, highly absorbing sample of the material being analyzed
对光漂移的评估可以采用高浓度,高吸收率的样品来分析
Stray Light Experiment: Fluorescent light, USB2000-VIS/NIR. Blocking provided by two OG550 filters (OD>7). Stray light was about 0.4 counts. Stray light software correction was set to 17. This gave closest fit to zero stray light. When the correction is "on" stray light was reduced to below the noise level of the measurement
光漂移实验:荧光灯,USB4000-VIS/NIR. 两个OG550滤光片(OD>7).光漂移大约0.4 counts.光漂移软件校正设置在17,这样最接近零漂移.当校正选项开启时光漂移减少到噪声检测水平以下.
Data Normalization 数据标准化
Spectrometer Signal Comes From:
Collection Optics Efficiency
Optical Fiber Attenuation
Optical Bench Efficiency
Fiber Mode Structure
Detector Response & linearity
Detector Interference Pattern
Pixel Fixed Pattern Noise
Amplifier electrical zero
and of course… the sample spectra
光谱信号受下列因素影响:
光采集效率
光纤响应曲线
光学平台的效率
光纤模式
探测器响应及线性度
探测器干扰
像素固定模式噪声
放大电路电子零点
采样光谱
Data Normalization 数据标准化
S(i) - D(i)
R(i) - D(i)
T(i) =
S(i) - D(i)
R(i) - D(i)
A(i) = -log
I(i) = B(i)
S(i) - D(i)
R(i) - D(i)
Percent of Reference Spectra
参考光谱的比例
Transmission, Reflection
透射,反射
Optical Density or Absorbance
光强或吸收 -log of T
Relative Irradiance
相对辐射量
Energy 能量
Choice of Reference for Transmission
透射参考值的选择
Choice of Reference for Reflection
反射测量的参考值选择
Choice of Reference for Energy
光强测量参考值
Normalized Data
数据标准化
T
I
A
Who Needs Data Normalization
谁需要数据标准化
Everybody! 任何人
Scope Mode data can't be compared with any other spectrometer
Scope模式的数据不能和其它任何光谱仪数据相比
Scope Mode is not quantitative if used at more than one wavelength (or pixel)
Scope模式在超过一个波长(像素)时不能量化
Spectra are distorted in shape 光谱形状被扭曲
Except 除非
If you know all of that and you calibrate scope mode to some external set of standards, for example when setting up a fluorescence signal vs concentration curve.
在Scope模式下为所有标准校正,例如荧光信号相对于浓度曲线的校正.
Spectral Processing
光谱处理
Correct Sample channel for drift observed in reference channel
为参考通道的漂移校正样品通道
Internal Standards (wavelength ratios)
内部标准(波长比率)
Math applied to whole spectra sets:
应用到整个光谱分析的数学知识
Subtracting a background spectra 去除背景噪声光谱
Dividing by a reference spectra 除以参考光谱值
Derivativation (1st and 2nd derivative of Amplitude vs Wavelength)
求导(振幅对波长的一阶和二阶导数)
Deconvolution & Curve Fitting 去卷积和曲线拟合
Gaussian, Lorentzian 高斯,拉葛朗日插值
FFT, peak integration, interpolation, etc.
FFT,峰值积分,插值
Sample样品光谱
Reference参考光谱
Reference参考光谱
sample样品光谱
For Absorbance对吸收率测量,Scor(l) = Sobs(l) - R(l)
For Transmission对投射率测量,Scor(l) = Sobs(l) / R(l)
Monitor lamp drift
监测灯光的漂移
Monitor lamp & Sample drift
监测灯光和样品的漂移
Reference Monitoring Options
参考值监测选项
Any channel of spectrometer can use any other channel (or itself) as the reference monitor
每个通道的光谱仪可以利用其它通道的光谱仪作为参考光谱监视器
Three Options 三种选择
Single Point 单点
Wavelength by Wavelength逐波长
Integrated band 积分带
Who Needs Reference Monitoring
谁需要监视参考光谱
Any long term experiment when taking a new reference is not an option
任何没有及时获取参考光谱的长时间实验
Any experiment that looks at the difference between two samples
任何在两次测量有差异的实验中
Any experiment that has samples with interfering backgrounds
样品有背景干扰的实验
Trade-offs – Answer is only as good as the reference monitoring data. If its not necessary, don't use it.
非必要,无需用
Wavelength Ratios & Baseline Corrections – Internal references
波长比率和基准线校正-内部参考
Simple Algorithms can be written in the Time Series Analysis Function
样品光谱可以在时间序列分析功能中计算
Full Spectral Math Functions are available in OOIBase32 Platinum
在OOIBase32白金版中有所有关于光谱计算的数学函数
AddConstant
Adds a constant to all points in a spectrum
每个点加上一个常量
AddSpectra
Adds two spectra, point-by-point
逐点相加两个光谱
Average
Calculates the average value of a spectrum
计算光谱的平均值
AverageWindow
Calculates the average value of a spectral window
计算一段光谱的平均值
Boxcar
Boxcar smooths (non-weighted averaging) a spectrum
对一个光谱进行平滑处理
CalculateAbsorbance
Calculates absorbance for a set of spectra
计算一系列光谱的吸收率
CalculateIrradiance
Calculates relative irradiance for a set of spectra
计算一系列光谱的相对辐射强度
CalculateTransmission
Calculates transmission for a set of spectra
计算一系列光谱的投射率
Derivative
Calculates the derivative of a spectrum
计算光谱的导数
DivideConstant
Divides a constant into all points in a spectrum
光谱每个点除以一个常量
DivideSpectra
Divides two spectra, point-by-point
两光谱逐点相除
GetMaxValue
Calculates the maximum value of a spectrum
计算光谱的极大值
GetMinValue
Calculates the minimum value of a spectrum
计算光谱的极小值
Integrate
Integrates an entire spectrum
在整个光谱上积分
IntegrateWindow
Integrates a region of a spectrum
在一段光谱上积分
Interpolate
Interpolates (using cubic splines) a spectrum to evenly-spaced wavelength intervals
在光谱上进行等距插值(三次样条)
Log10Spectrum
Computes the log10 of a complete spectrum
光谱取对数
MultiplyConstant
Multiplies a constant to all points in a spectrum
在整个光谱上乘以一个常数
MultiplySpectra
Multiplies two spectra, point-by-point两个光谱逐点相乘
PolynomialRegression
Computes a polynomial regression on a set of data
计算多项式回归系数
SubtractConstant
Subtracts a constant to all points in a spectrum
在整个光谱上逐点减去一个常数
SubtractSpectra
Subtracts two spectra, point-by-point两个光谱逐点相减
Sum
Sums all points in a spectrum
整个光谱求和
SumWindow
Sums a region of a spectrum
一段光谱求和
Spectral Modeling
光谱建模
Chromaticity (L* a* b*)色度
Provides a scale related to the appearance of the color to a human eye
提供人眼区分颜色的梯度系统
Beer's Law 比尔法则
Used to calculate concentration of an analyte
计算浓度
Principle Component Analysis 主成分分析
Used to analyze samples with more than one analyte
分析多种混合样品
PLS and other Chemometric Tools 最小平方法及其它
Used to analyze more complex mistures of signals
用于分析更多复杂的混合信号
Spectral Modeling
光谱建模
Neural Nets 神经网络
A training approach to build an empirical model
通过训练的方法建立实验模型
Radiometry 幅射线测定
Measures radiant energy
测量辐射能量
Thin film analysis 薄膜分析
Metrology of film thickness from interference spectra
通过干涉光谱分析薄膜厚度
Scattering & particle size analysis 散射和粒子大小分析
Analyzing particle size from scattering spectra
通过散射光谱分析粒子大小
Thanks
谢谢!
·上一篇：光电子物理基础
·下一篇：可提供之合作研究及服务项目